25 mev proton events observed by the high energy ... code 290-17, ... 101 102 stereo a 14-24 mev...

Solar PhysicsDOI: 10.1007/•••••-•••-•••-••••-•

> 25 MeV Proton Events Observed by the High

Energy Telescopes on the STEREO A and B

Spacecraft and/or at the Earth During the First

∼Seven Years of the STEREO Mission

I. G. Richardson1,2· T. T. von Rosenvinge1

·

H. V. Cane3· E. R. Christian1

·

C. M. S. Cohen4· A. W. Labrador4

·

R. A. Leske4· R. A. Mewaldt4

·

M. E. Wiedenbeck5· E. C. Stone4

c© Springer ••••

Abstract Using observations from the High Energy Telescopes (HETs) on theSTEREO A and B spacecraft, and similar observations from near-Earth space-craft, we summarize the properties of more than 200 individual > 25 MeV solarproton events, some detected by multiple spacecraft, that occurred from the be-ginning of the STEREO mission in October 2006 to December 2013, and providea catalog of these events and their solar sources and associations. Longitudinaldependencies of the electron and proton peak intensities and delays to onsetand peak intensity relative to the solar event have been examined for 25 three-spacecraft particle events. Expressed as Gaussians, peak intensities fall off withlongitude with σ = 47 ± 14o for 0.7–4 MeV electrons, and σ = 43± 13o for 14–24 MeV protons. Several particle events are discussed in more detail, includingone on 3 November 2011, in which ∼ 25 MeV protons filled the inner heliospherewithin 90 minutes of the solar event, and another on 7 March 2012, in which we

1NASA/Goddard Space Flight Center, Greenbelt, Maryland,USA 20771email: [email protected];[email protected]; [email protected] and Department of Astronomy, University ofMaryland, College Park, Maryland, USA 207423Bruny Island Radio Spectrometer, Bruny Island, Tasmania,Australiaemail: [email protected] Code 290-17, California Institute of Technology,Pasadena, CA, USA 91125email: [email protected]; [email protected];[email protected]; [email protected];[email protected];5Jet Propulsion Laboratory, California Institute ofTechnology, Pasadena, CA, USA 91109email: [email protected]© 2014. All rights reserved.

SOLA: multiscsep6_fin.tex; 24 March 2014; 14:42; p. 1

I. G. Richardson et al.

demonstrate that the first of two CMEs that erupted from an active region within

∼ 1 hour was associated with particle acceleration. Comparing the current Solar

Cycle 24 with the previous cycle, the first > 25 MeV proton event was detected

at Earth in the current solar cycle around one year after smoothed sunspot

minimum, compared with a delay of only two months in Cycle 23. Otherwise,

SEP event occurrence rates were reasonably similar during the rising phases of

Cycles 23 and 24. However, the rate declined in 2013, reflecting the decline in

sunspot number since the peak in the northern-hemispheric sunspot number in

November 2011. Observations in late 2013 suggest that the rate may be rising

again in association with an increase in the southern sunspot number.

Keywords: Solar energetic particles, STEREO, SOHO

1. Introduction

With the launch of the STEREO A (“Ahead”) and B (“Behind”) spacecraft on

26 October 2006 into heliocentric orbits advancing ahead of or lagging the Earth

in its orbit, respectively, solar energetic particle (SEP) events can be observed at

multiple locations near 1 AU. Even when closely separated shortly after launch,

interesting differences in the particle intensities at STEREO A and B and near-

Earth spacecraft were observed during the final large SEP events of Solar Cycle

23 in December 2006 (von Rosenvinge et al., 2009). Separating from the Earth

by ∼ 22o/year, STEREO A and B were 180o apart above the west and east

limbs of the Sun, respectively, as viewed from the Earth on 6 February 2011,

allowing observations of the complete solar surface to be made for the first time.

In this paper, we focus on observations of 14–41 MeV protons and 0.3–

4 MeV electrons made by the High Energy Telescopes (HETs) on the STEREO

spacecraft (von Rosenvinge et al., 2008) during the first ∼seven years of the

STEREO mission. By combining these observations with similar observations

from the EPHIN (Muller-Mellin et al., 1995) and ERNE (Torsti et al., 1995)

instruments on the SOHO spacecraft in orbit around the L1 point upstream of

Earth, we have compiled a catalog of 209 individual > 25 MeV proton events

that were observed at one or multiple spacecraft from the start of the STEREO

mission up to the end of 2013, when the STEREO spacecraft were separated

by ∼ 60o on the far side of the Sun relative to Earth. We then summarize the

general features of a subset of events observed at all three locations, including

how the particle onset delay and peak intensity depend on the location of the

associated solar event relative to the observing spacecraft. The results provide

constraints for models of particle transport processes at/near the Sun or in the

solar wind and suggest a formula for the SEP intensity as a function of the speed

of the associated coronal mass ejection and location of the solar event relative

to the observer which is tested using a sample of more than 500 events in Cycles

23 and 24.


Multi-spacecraft SEP events

Cycle 23Cycle 24

0

50

100

10-410-310-210-1100101102

STEREO A 14-24 MeV Protons

10-410-310-210-1100101102

Inte

nsity (

Me

V s

cm

**2

sr)

**-1

10-410-310-210-1100101102

2008 2010 2012 2014

Ò�

STEREO B 14-24 MeV Protons

SOHO/ERNE 14-24 MeV Protons

Sunspot NumberSmoothed

Figure 1. Intensities of 14–24 MeV protons (1-hr averages) observed by the HET instrumentson STEREOs A (Ahead) and B (Behind) (second and bottom panels, respectively) and theERNE instrument on SOHO (third panel), from October 2006 to December 2013. The toppanel shows the monthly and smoothed international sunspot numbers.

2. Observations

In this study, we use energetic particle observations made by the HET in-struments on STEREOs A and B and the ERNE and EPHIN instrumentson SOHO from STEREO launch in October 2006 until December 2013. Forprotons, we focus on two energy ranges, 14–24 and 24–41 MeV, defined by acombination of HET energy channels to enhance the counting statistics. HETdata are available from http://www.srl.caltech.edu/STEREO/. Energy channels36–42 and 44–50 of the ERNE High Energy Detector closely match these en-ergy ranges (13.8–24.2 and 24.1–40.5 MeV, respectively); the ERNE data wereobtained from the Space Research Laboratory, University of Turku (http://www.srl.utu.fi/erne data/). EPHIN has proton channels at 7.8–25 MeV and 25–53 MeV; the EPHIN data were obtained from the University of Kiel (http://www.ieap.uni-kiel.de/et/ag-heber/costep/). For electrons, we use intensities at0.7–4 MeV from the STEREO A/B HETs, and at 0.67–3 MeV from EPHIN.

Figure 1 gives an overview of 14–24 MeV proton intensities observed by theSTEREO A and B HETs (second and bottom panels) and ERNE (third panel),from October 2006 to December 2013. The top panel shows the monthly averageand smoothed sunspot number, indicating that this interval extended from thelate decay phase of Solar Cycle 23 to the first ∼five years of Cycle 24 (smoothedsunspot minimum was in December 2008). Following the intense particle eventsin December 2006 (e.g., von Rosenvinge et al., 2009), there were few SEP events



0

50

100

10-410-310-210-1100101102

STEREO A 14-24 MeV Protons

10-410-310-210-1100101102

Inte

nsity (

Me

V s

cm

**2

sr)

**-1

10-410-310-210-1100101102

2011

J F M A M J J A S O N D J

2012

F M A M J J A S O N D J

2013

F M A M J J A S O N D J

2014

Ò�

STEREO B 14-24 MeV Protons

SOHO/ERNE 14-24 MeV Protons

Sunspot NumberSmoothed

Figure 2. As in Figure 1 for January 2011 to December 2013.

extending into this energy range during the extended sunspot minimum betweenthese cycles. The SEP occurrence rate finally increased in December 2009, aroundone year into Cycle 24. By this time, STEREO A was 64o ahead of the Earthwhile STEREO B was lagging the Earth by 67o. Note that while the SEP eventsin December 2009–February 2010 were relatively weak, those on 22 December2009 and 17 January 2010 were observed at all three spacecraft. This increase inactivity was temporary, however. Few SEP events were observed during Marchto August 2010, when activity increased once again, but also temporarily. SEPevents were observed more persistently from around February 2011. Thus, thefirst ∼ 2 years of Cycle 24 were characterized by brief episodes of > 14 MeVproton events during an otherwise quiet period. Interestingly, the intervals be-tween these episodes of activity are both ∼ 6–7 months. Quasi-periodicities of“∼ 150 days” have been reported for energetic solar phenomena in some othercycles, for example by Rieger et al. (1984), Lean (1990) (who identified periods of∼ 130–185 days in a survey of multiple solar cycles), Cane, Richardson and vonRosenvinge (1998), Dalla et al. (2001), Richardson and Cane (2005), Richardsonand Cane (2010), and references therein, so the rise of Cycle 24 may also showevidence of a similar phenomenon.

Figure 2 shows an expanded view of observations from January 2011 (whenSTEREO A was 86o ahead of the Earth and STEREO B was behind by 91o)to December 2013 in the format of Figure 1. By the end of December 2013,STEREO A was 150o ahead of the Earth, STEREO B was behind by 152o, andthe STEREO spacecraft were separated by 58o. Despite the large separation be-tween the spacecraft, overall, the particle intensity–time profiles are remarkably



similar, with intervals of higher intensity that include contributions from thesame solar events, as will be discussed below, interspersed with quieter intervalsthat are also evident at all locations.

The inter-calibration between the various instruments used in this study canbe checked over a wide dynamic range during the events in December 2006, whenthe STEREO spacecraft were still close to Earth. Figure 3 compares intensitiesmeasured during December 2006 by various instruments. Figure 3(a) shows one-hour averages of the HET A 13.6–23.8 MeV proton intensity plotted against theHET B intensity at the same energy. As might be expected, the HET A andHET B intensities are highly correlated (cc = 0.991), with similar intensities atboth spacecraft (I(B) = 1.006I(A)0.992; the red line in each panel in Figure 3indicates equality between the intensities). Note that STEREO B was inverted atthis time so that HET B was viewing perpendicular to the nominal Parker spiraldirection whereas the STEREO A HET was viewing along the spiral direction.Differences in the intensities measured by the HETs might be expected at timesof significantly anisotropic particle distributions, such as those discussed by vonRosenvinge et al. (2009), but evidently the observed intensities were generallycomparable despite the different spacecraft configurations.

Figure 3(b) shows the 13.8–24.2 MeV proton intensity from the SOHO/ERNEHED plotted against the same HET A intensity. The intensities are corre-lated, though with the ERNE intensity ∼ 1.5 times the HET A intensity, until∼ 1 (MeV s cm2 sr)−1 when ERNE saturates. Figure 3(c) shows the 24.1–40.5 MeV proton intensity from ERNE plotted vs. the 23.8–40.5 MeV protonintensity from HET A, again showing close agreement until ERNE saturatesabove ∼ 0.1 (MeV s cm2 sr)−1. Figure 3(d) shows the 7.8–25 MeV proton inten-sity from EPHIN vs. the HET A 13.6–23.8 MeV proton intensity. The EPHINintensity is correlated with, but ∼ 1.7 times larger than the HET A intensity at aslightly higher energy. Hence, the EPHIN data, with this conversion factor, maybe used to provide an estimate of the intensity in the HET A energy range whenERNE is saturated or data are not available. Similarly, Figure 3(e) shows thatthe 25–53 MeV EPHIN proton intensity is around a half of the HET A 23.8–40.5 MeV intensity. Finally, Figures 3(f) and (g) compare electron intensities at0.7–4 MeV from HET A and 0.67–3 MeV from EPHIN, (f) showing the observedintensities, and (g) with a background of 0.015 (MeV s cm2 sr)−1 removed fromHET A. With this correction, the HET A and EPHIN electron intensities arewell correlated, but the EPHIN intensity is a factor of ∼ 14 above the HET Aintensity. (The reason for this relatively large conversion factor, which appearsto be associated with the HETs, since the EPHIN calibration is consistent withother instruments, is under investigation at the time of writing.) A similaranalysis for HET B suggests a similar conversion factor and a background of0.02 (MeV s cm2 sr)−1. The few points lying off the main distribution in (f)and (g) are associated with high electron intensities observed when STEREO Acrossed the geomagnetic tail.

SOHO makes a 180o roll approximately every three months (because of animmovable high gain antenna) causing the fields of view of EPHIN and ERNEto change direction. We have not attempted to make any correction to theEPHIN and ERNE intensities when this occurs. One reason is the STEREO



a

10-4

10-3

10-2

10-1

100

101

102

10-4

10-3

10-2

10-1

100

101

102

B 1

3.6

-23.8

MeV

HA 13.6-23.8 MeV H

cc=0.991ln(y)=0.992*ln(x)+0.0059

b

10-4

10-3

10-2

10-1

100

101

10-4

10-3

10-2

10-1

100

101

ER

NE

13.8

-24.2

MeV

H

A 13.6-23.8 MeV H

cc=0.987ln(y)=0.885*ln(x)-0.21

c

10-4

10-3

10-2

10-1

100

101

10-4

10-3

10-2

10-1

100

101

ER

NE

24.1

-40.5

MeV

H

A 23.8-40.5 MeV H

cc=0.982ln(y)=0.847*ln(x)-0.85

d

10-4

10-3

10-2

10-1

100

101

102

10-4

10-3

10-2

10-1

100

101

102

EP

HIN

7.8

-25 M

eV

H

A 13.6-23.8 MeV H

cc=0.991ln(y)=1.02*ln(x)+0.57

e

10-4

10-3

10-2

10-1

100

101

10-4

10-3

10-2

10-1

100

101

EP

HIN

25-5

3 M

eV

H

A 23.8-40.5 MeV H

cc=0.991ln(y)=0.953*ln(x)-0.75

f

10-2

10-1

100

101

102

103

10-2

10-1

100

101

102

103

EP

HIN

0.6

7-3

Me

V e

-

A 0.7-4 MeV e

cc=0.948ln(y)=1.16*ln(x)+2.29

g

10-2

10-1

100

101

102

103

10-2

10-1

100

101

102

103

EP

HIN

0.6

7-3

Me

V e

-

A 0.7-4 MeV-0.015

cc=0.976ln(y)=0.928*ln(x)+2.64

Figure 3. Comparison of the STEREO HETs, SOHO/ERNE E, and SOHO/EPHIN protonand electron intensities (in (MeV s cm2 sr)−1) using observations during December 2006. (a)HET B vs. HET A for 13.6–23.8 MeV protons; (b) ERNE 13.8–24.2 MeV protons vs. HET A13.6-23.8 MeV protons; (c) ERNE 24.1–40.5 MeV protons vs. HET A 23.8–40.5 MeV protons;(d) EPHIN 7.8–25 MeV protons vs. HET A 13.6–23.8 MeV protons; (e) EPHIN 25-53 MeVprotons vs. HET A 23.8–40.5 MeV protons; (f) EPHIN 0.67–3.0 MeV electrons vs. HET A0.7–4.0 MeV electrons, and (g) with a HET background of 0.015 (MeV s cm2 sr)−1 subtracted.Black lines are least squares fits to the data, while red lines indicate equal values.



observations in Figure 3(a), made when STEREO B was inverted, suggest thatthe correction might be small. Furthermore, when the STEREO and SOHOspacecraft are well separated, they measure different particle populations withdifferent anisotropies. An accurate correction would require information on theparticle pitch angle distribution and IMF at each spacecraft. Thus, in this paper,we use the observed particle intensities at each spacecraft without any correction(other than the calibrations suggested by Figure 3) with the expectation thatthese are accurate to within a factor of ∼ 2.

2.1. Event Catalog

In past studies (e.g., Cane, Reames and von Rosenvinge, 1988; Cane, Richardsonand von Rosenvinge, 2010a), we have documented the properties of particleevents that include ∼ 25 MeV protons. In particular, Cane, Richardson andvon Rosenvinge (2010a) discussed 280 such events that occurred in 1997–2006(see also Vainio et al., 2013). To compile a similar catalog of SEP events at theSTEREO spacecraft and/or at the Earth that include protons of such energies,we have examined proton and electron intensity–time data from each spacecraft,as described above, with integration periods of 1 minute to 1 hour. At suchenergies, the SEP intensity is generally not dominated by locally-acceleratedions associated with interplanetary shocks, and peak intensities typically occurearly in the event, providing information on particle acceleration close to theSun.

To identify the associated solar event and its location, we have examined:movies from the SOHO EIT instrument (Delaboudiniere et al., 1995), the EUVIinstruments on the STEREO spacecraft (Wulser et al., 2004), and the AIA onthe Solar Dynamics Observatory (Lemen et al., 2012); observations of coro-nal mass ejections from the LASCO coronagraphs on SOHO (Brueckner etal., 1995) and the SECCHI COR1 and COR2 coronagraphs on the STEREOspacecraft (Howard et al., 2008); Hα and GOES soft X-ray flare reports (limitedto front side or near-limb events), and observations of solar radio emissions madeby the WAVES instrument on WIND (http://lep694.gsfc.nasa.gov/waves/) andthe SWAVES instruments on the STEREO spacecraft (http://swaves.gsfc.nasa.gov/). Observations by WAVES and SWAVES of the type III radio emissions thataccompany nearly every SEP event (e.g., Cane, Erickson and Prestage (2002),Cane and Erickson (2003), Cane, Richardson and von Rosenvinge (2010a); seealso below) are of particular value since the intensity and degree of occultationat higher frequencies by the limb of the Sun observed at the three locations canhelp to confirm the location of the SEP source at the Sun. We use the onset timesof the energetic particles at the three spacecraft, in particular that showing themost prompt increase, to infer an approximate time for the event at the Sun.Particle event onset times are estimated by inspection of intensity–time profilesaround event onset using suitable intensity/time scales and data-averaging inter-vals to estimate when the intensity rises above the range of intensity fluctuationsin the pre-event background. We then examine the solar and radio observationsfor evidence of, for example, a flare, eruption in EIT, EUV or EUVI, a coronalmass ejection (CME), and type III and/or type II radio emissions (type II radio



emissions were obtained from the WIND WAVES/STEREO SWAVES list athttp://lep694.gsfc.nasa.gov/waves/data products.html) that is consistent with theSEP event onset. For every SEP event including 25 MeV protons identified sofar, it is possible to make an unambiguous solar event association. In particular,the STEREO spacecraft can, for the first time, provide imaging evidence that anevent originated on the far side of the Sun from Earth. We also examine SEP dataover a range of energies and solar wind observations at the STEREO spacecraftand at Earth to identify and remove any intensity increases that are likely tobe associated with the passage of shocks, or are spatial modulations that, forexample, have no velocity dispersion or are related to solar wind structures. Suchfeatures are also not usually associated with appropriately timed solar activity.

For most of the events since August 2010, we have compared the inferredsolar source longitudes with the propagation direction of the related CMEsobtained by triangulation of the STEREO and LASCO coronagraph observa-tions included in the “Space Weather Database Of Notifications, Knowledge,Information (DONKI)” (http://kauai.ccmc.gsfc.nasa.gov/DONKI/) developed atthe Community Coordinated Modeling Center. As would be expected, the CMEpropagation longitudes are usually consistent with the source longitudes (in 82%of cases, the directions differed by ≤ 20o) suggesting that the source identifica-tions are likely to be correct. For 5% of the events, the directions differed by over60o. These cases were then re-checked, including consideration of the particle,radio and solar data, and the source location was either reconciled with the CMEdirection, or the original location was retained if the supporting evidence wasmore compelling.

The 209 unique > 25 MeV proton events identified up to December 2013are listed in Tables 1–7 (a subset of these events are also considered by Larioet al., 2013). The first two columns give the times (date and hour of day) ofthe associated solar event and the particle event onset at the spacecraft withthe most prompt particle arrival. In most cases, these times differ by only anhour or two, but in some cases, the delay is larger, or the solar event is onthe day preceding the SEP onset. The next three pairs of columns show, forSTEREO B, the Earth (i.e., SOHO) and STEREO A, the heliolongitude of thesolar event relative to the observing spacecraft (a negative value indicates thatthe event is east of the spacecraft) and the peak intensity for ∼ 25 MeV protonsin (MeV s cm2 sr)−1. The intensity is generally estimated within the first 24hours or so of the event, so that a higher peak in association with passage of aninterplanetary shock (generally rare at these energies) is not taken as the “peak”of the event. In some cases, a slowly rising, extended particle event is observedwithout an in-situ shock, and peak intensity may be estimated at a later time.“BG” indicates that there is a high background from a preceding event thatmight obscure the event at that spacecraft, while “...” means that the event wasnot detected at ∼ 25 MeV above a threshold of ∼ 10−4 (MeV s cm2 sr)−1 .

The following column shows the GOES soft (1–8 A) X-ray flare intensity, forfront-side or near-limb events. The presence of type III radio emissions below20 MHz in the daily summary plots of WIND/WAVES or STEREO SWAVESdata is indicated by ‘1’ in the next column, similarly in the next column for anevent in the list of “possible” type II emissions observed by WIND/WAVES or


Multi-sp

acecra

ftSE

Pev

ents

Table 1. > 25 MeV Proton Events at 1 AU During the STEREO Mission

Solar Event SEP Event Flare–B I(B) Flare–Earth I(Earth) Flare–A I(A) GOES Type Type CME dA CME V(UT) (UT) (o) (o) (o) X-Ray III II (o) km s−1

2006 Dec 5, 10 Dec 5, 12 -68 0.04 -68 0.04 -68 0.04 X9.0 1 2 DG DG2006 Dec 6, 18 Dec 6, 19 -64 BG -64 BG -64 BG X6.5 1 2 360 DG2006 Dec 13, 02 Dec 13, 02 23 10 23 10 23 10 X3.4 1 2 360 17742006 Dec 14, 21 Dec 14, 22 46 0.5 46 0.5 46 0.5 X1.5 1 2 360 10422007 May 19, 12 May 19, 14 08 0.0004 05 0.0009 -01 0.0003 B9.5 1 1 106 9582007 May 23, 07 May 23, 07 63 0.0003 60 0.0003 54 0.0003 B5.3 1 0 90 6792007 Dec 31, 00 Dec 31, 02 -85 ... -108 0.0001 -129 ... C8.3 1 1 164 9952009 May 5, 08 May 5, 10 -90 ... -137 ... 175 0.0006 ... 1 1 66 B 595 B2009 Dec 22, 04 Dec 22, 05 113 0.0008 46 0.0005 -18 0.0002 C7.2 1 0 47 3182010 Jan 17, 03 Jan 17, 04 -50 0.0003 -120 0.0003 176 0.003 ... 1 1 126 3502010 Feb 6, 07 Feb 6, 08 53 0.0003 -18 ... -83 ... C4.0 1 0 56 2402010 Feb 7, 02 Feb 7, 03 61 0.0003 -10 ... -75 ... M1.3 1 0 360 4212010 Feb 12, 11 Feb 12, 12 60 0.001 -11 0.0003 -76 ... M8.3 1 0 360 5092010 Jun 12, 00 Jun 12, 01 113 0.0001 43 0.007 -31 0.0002 M2.0 1 1 119 4862010 Aug 1, 07 Aug 1, 12 35 0.02 -36 0.0005 -114 ... C3.2 0 2 360 8502010 Aug 7, 17 Aug 7, 19 38 0.08 -34 0.005 -113 ... M1.0 1 2 360 8712010 Aug 14, 09 Aug 14, 10 126 0.004 54 0.16 -26 0.001 C4.4 1 0 360 12052010 Aug 18, 05 Aug 18, 06 172 0.001 100 0.03 20 0.008 C4.5 1 2 184 14712010 Aug 31, 02 Aug 31, 03 -141 ... 145 ... 64 0.0002 ... 1 0 33 3192010 Aug 31, 21 Aug 31, 22 -141 0.0001 145 0.005 64 0.008 ... 1 1 360 13042010 Sep 4, 14 Sep 4, 20 144 ... 70 0.0001 -11 ... B2.5 1 0 130 5672010 Sep 8, 23 Sep 8, 24 167 0.0003 92 0.003 10 0.003 C3.3 1 1 147 8182010 Dec 31, 04 Dec 31, 05 159 ... 69 0.0001 -17 ... C1.3 1 0 45 3632011 Jan 27, 08 Jan 27, 10 169 ... 76 0.0001 -10 ... B6.6 1 0 52 3162011 Jan 27, 11 Jan 27, 14 178 ... 85 0.0001 -1 ... C1.2 1 1 43 3492011 Jan 28, 00 Jan 28, 02 -177 ... 90 0.05 04 0.02 M1.3 1 1 119 6062011 Jan 31, 16 Jan 31, 20 -136 ... 131 ... 45 0.0002 ... 1 1 150 7132011 Feb 13, 17 Feb 13, 18 90 0.0004 -04 0.00015 -91 ... M6.6 1 1 276 3732011 Feb 14, 17 Feb 14, 18 112 0.001 18 0.0003 -69 ... M2.2 1 0 360 3262011 Feb 15, 01 Feb 15, 02 112 0.08 18 0.021 -69 ... X2.2 1 2 360 6692011 Feb 24, 07 Feb 24, 08 15 0.0006 -102 ... -171 ... M3.5 1 0 158 11862011 Mar 4, 14 Mar 4, 15 -85 ... 180 ... 90 0.0003 ... 1 0 60 B 297 B2011 Mar 7, 13 Mar 7, 15 74 0.04 -21 0.0004 -109 ... M1.9 1 1 261 6982011 Mar 7, 19 Mar 7, 21 148 0.4 53 0.5 -35 0.006 M3.7 1 2 360 2125

SOLA:multiscsep6_fin.tex;24March2014;14:42;p.9

I.G

.R

ichard

son

etal.



2011 Mar 16, 17 Mar 16, 22 -167 ... 72 0.004 -16 0.0002 C3.7 1 0 184 6822011 Mar 21, 02 Mar 21, 03 -127 0.0008 138 0.11 50 12 ... 1 2 360 13412011 Mar 29, 20 Mar 30, 04 -25 0.01 -120 ... 151 0.003 ... 0 0 > 195 12642011 Apr 3, 05 Apr 3, 06 170 ... 75 0.0008 -14 ... C1.2 1 0 DG DG2011 Apr 4, 03 Apr 4, 12 165 ... 70 0.001 -19 ... B8.6 1 0 DG DG2011 Apr 21, 00 Apr 21, 02 -105 ... 160 0.0007 70 0.001 ... 1 0 111 4752011 May 9, 20 May 10, 00 0 0.002 -94 ... 174 ... C5.4 1 2 292 13182011 May 11, 02 May 11, 04 147 ... 53 0.004 -39 ... B8.1 1 0 225 7452011 May 19, 03 May 19, 06 154 ... 60 ... -33 0.0001 B3.0 1 0 132 6262011 May 29, 20 May 29, 22 25 0.01 -68 ... -162 ... C8.7 1 2 186 14072011 Jun 2, 07 Jun 2, 09 68 0.0006 -25 ... -119 ... C3.7 1 1 360 9762011 Jun 4, 06 Jun 4, 07 -112 ... 155 0.002 60 0.1 ... 1 2 360 14072011 Jun 4, 22 Jun 4, 22 -100 0.08 165 0.04 70 13 ... 1 2 360 24252011 Jun 7, 06 Jun 7, 07 147 BG 54 0.7 -41 BG M2.5 1 2 360 12552011 Jun 29, 00 Jun 29, 00 169 ... 77 0.0001 -20 ... B7.3 1 0 122 4812011 Jul 11, 10 Jul 11, 12 86 ... -06 0.0001 -104 ... C2.6 1 0 53 2662011 Jul 26, 08 Jul 26, 10 -82 0.0004 -175 ... 85 0.01 ... 1 1 360 3822011 Aug 2, 05 Aug 2, 07 108 0.0008 15 0.04 -85 ... M1.4 1 1 268 7122011 Aug 3, 13 Aug 3, 14 123 ... 30 0.007 -70 ... M6.0 1 0 360 6102011 Aug 4, 03 Aug 4, 05 129 0.004 36 1.1 -64 0.005 M9.3 1 2 360 13152011 Aug 8, 18 Aug 8, 19 154 ... 61 0.05 -39 ... M3.5 1 2 237 13432011 Aug 9, 07 Aug 9, 08 162 ... 69 0.5 -31 0.002 X6.9 1 1 360 16102011 Aug 23, 22 Aug 24, 06 -85 ... -179 ... 70 0.002 ... 0 0 360 3122011 Sep 3, 02 Sep 3, 03 -75 ... 170 0.0002 68 0.0003 ... 1 0 105 3252011 Sep 4, 04 Sep 4, 05 162 ... 67 0.002 -35 ... C9.0 0 0 53 2622011 Sep 4, 23 Sep 5, 02 -165 ... 100 0.0006 -03 0.0002 C7.9 1 0 146 6222011 Sep 6, 01 Sep 6, 02 102 0.003 7 0.03 -95 ... M5.3 1 2 360 7822011 Sep 6, 22 Sep 6, 23 113 0.005 18 0.1 -84 0.0002 X2.1 1 2 360 5752011 Sep 7, 18 Sep 7, 20 -50 BG -145 BG 112 0.004 ... 1 1 188 9242011 Sep 8, 22 Sep 9, 00 -122 BG 143 BG 40 0.003 ... 1 1 281 9832011 Sep 13, 12 Sep 13, 12 -161 ... 103 0.0003 0 ... ... 1 0 199 7462011 Sep 21, 22 Sep 21, 23 -155 ... 109 0.002 05 0.0004 ... 1 0 255 10072011 Sep 22, 10 Sep 22, 10 18 10.0 -78 0.08 -179 0.04 X1.4 1 2 360 19052011 Sep 24, 09 Sep 24, 11 37 0.5 -60 BG -164 BG X1.9 1 2 145 1936


Multi-sp

acecra

ftSE

Pev

ents



2011 Oct 4, 12 Oct 4, 15 -50 0.2 -148 0.001 108 0.02 ... 1 0 360 11012011 Oct 14, 11 Oct 14, 15 -40 0.0002 -140 ... 115 0.0005 ... 1 0 241 8142011 Oct 20, 03 Oct 20, 04 -155 ... 105 0.0008 0 0.0002 M1.6 1 0 193 8932011 Oct 22, 10 Oct 22, 12 177 0.0003 77 0.006 -28 0.0008 M1.3 0 2 360 11052011 Nov 3, 22 Nov 3, 23 -50 0.06 -152 0.04 103 2.0 ... 1 2 360 9912011 Nov 9, 13 Nov 9, 14 83 0.002 -20 0.002 -126 ... M1.1 1 2 360 9072011 Nov 12, 18 Nov 12, 18 0 0.0002 -103 ... 151 0.0003 ... 1 0 192 6542011 Nov 13, 18 Nov 13, 22 -70 0.0003 -174 ... 80 0.001 ... 1 0 360 5962011 Nov 17, 22 Nov 17, 23 -20 0.01 -125 0.0005 129 0.001 ... 1 0 360 10412011 Nov 23, 15 Nov 23, 15 -60 ... -165 ... 90 0.0003 ... 1 0 95? 315?2011 Nov 26, 06 Nov 26, 08 154 0.002 48 0.33 -58 0.01 C1.2 1 2 360 9332011 Dec 3, 07 Dec 3, 08 -117 ... 136 DG 30 0.0003 ... 1 0 155 4542011 Dec 11, 13 Dec 11, 14 -76 ... 176 ... 70 0.0003 ... 1 0 32 A 312 A2011 Dec 17, 10 Dec 17, 16 -115 ... 136 0.0003 30 0.002 ... 1 0 247 9872011 Dec 24, 11 Dec 24, 14 -165 ... 85 0.0003 -22 ... C4.9 1 2 102 4472011 Dec 25, 18 Dec 25, 19 136 0.006 26 0.05 -81 ... M4.0 1 1 125 3662011 Dec 27, 04 Dec 27, 05 78 0.0006 -32 BG -139 ... C8.9 1 0 84 B 735 B2011 Dec 29, 16 Dec 29, 18 -142 ... 107 0.001 0 ... ... 1 0 171 7362012 Jan 2, 14 Jan 2, 16 -152 ... 97 0.015 -10 0.001 C2.4 1 1 360 11382012 Jan 12, 07 Jan 12, 10 -05 0.0003 -118 ... 135 0.0006 C2.5 0 0 360 8142012 Jan 16, 02 Jan 16, 08 45 0.0005 -68 0.0003 -175 ... C6.5 0 0 360 10602012 Jan 19, 13 Jan 19, 15 84 0.03 -30 0.0006 -137 ... M3.2 1 2 360 11202012 Jan 23, 04 Jan 23, 04 135 0.6 21 20 -87 0.1 M8.7 1 2 360 21752012 Jan 27, 17 Jan 27, 19 -175 BG 71 10 -37 0.3 X1.7 1 2 360 25082012 Feb 24, 04 Feb 24, 10 79 ... -38 0.0005 -147 ... ... 0 2 189 8002012 Feb 29, 09 Feb 29, 10 -118 ... 125 0.0005 16 0.005 ... 1 0 360 4462012 Mar 3, 18 Mar 4, 00 45 0.0002 -73 ... 178 ... C1.9 1 0 49 10782012 Mar 4, 10 Mar 4, 12 48 0.4 -61 0.02 -170 ... M2.0 1 2 360 13062012 Mar 5, 02 Mar 5, 04 57 0.3 -52 BG -161 0.001 X1.1 1 2 360 15312012 Mar 7, 00 Mar 7, 01 91 10.0 -27 10 -137 0.09 X5.4 1 2 360 26842012 Mar 13, 17 Mar 13, 18 179 BG 61 10 -48 BG M7.9 1 2 360 18842012 Mar 21, 07 Mar 21, 07 -81 BG 160 BG 50 0.8 ... 1 1 360 11782012 Mar 24, 00 Mar 24, 00 -50 0.03 -169 0.001 80 1.3 ... 1 2 360 11522012 Mar 26, 22 Mar 27, 00 -5 0.6 -124 BG 125 BG ... 1 1 360 13902012 Apr 5, 20 Apr 6, 00 148 ... 29 0.002 -82 ... C1.5 1 0 360 828


I.G

.R

ichard

son

etal.



2012 Apr 7, 16 Apr 7, 17 -80 ... 161 0.0005 50 0.02 ... 1 2 360 7652012 Apr 9, 12 Apr 9, 12 -171 ... 73 0.0006 -38 ... C3.9 1 1 360 9212012 Apr 15, 02 Apr 15, 03 29 0.002 -90 ... 158 0.001 C1.7 1 1 173 12202012 Apr 18, 02 Apr 18, 03 -39 ... -150 ... 90 0.0006 ... 1 0 > 271 3522012 Apr 18, 14 Apr 18, 15 -35 ... -154 0.0001 94 0.0004 ... 1 2 184 8402012 Apr 24, 07 Apr 24, 09 30 0.015 -88 0.0002 159 ... C3.7 1 0 190 4432012 Apr 27, 15 Apr 27, 18 -108 ... 133 ... 20 0.0003 ... 1 1 360 6812012 May 1, 16 May 1, 18 -59 ... -177 ... 70 0.0002 ... 1 0 200 7882012 May 7, 00 May 7, 03 142 ... 24 0.0002 -90 ... C1.3 1 0 124 A 625 A2012 May 14, 09 May 14, 10 158 ... 40 0.0003 -75 ... C2.5 1 0 48 5512012 May 17, 01 May 17, 02 -166 0.001 76 0.6 -39 0.002 M5.1 1 2 360 15822012 May 26, 20 May 26, 21 -127 0.001 116 0.03 00 0.2 ... 1 2 360 19662012 Jun 1, 22 Jun 1, 22 -162 ... 81 0.0006 -35 ... C3.3 1 0 175 6302012 Jun 3, 17 Jun 3, 19 79 0.01 -38 ... -154 ... M3.3 1 0 180 6052012 Jun 8, 07 Jun 8, 07 157 ... 40 0.0001 -77 ... C4.8 0 0 9 3292012 Jun 12, 05 Jun 12, 06 -109 ... 135 0.0001 18 0.001 ... 1 0 195 8642012 Jun 14, 12 Jun 14, 15 111 0.001 -06 0.002 -123 ... M1.9 1 0 360 9872012 Jun 28, 07 Jun 28, 08 -65 ... 179 ... 60 0.002 ... 1 0 360 7282012 Jun 28, 16 Jun 28, 16 71 0.001 -45 ... -164 ... M2.4 1 0 104 7142012 Jul 2, 08 Jul 2, 09 -10 0.001 -126 0.0001 115 0.01 ... 1 2 360 17042012 Jul 4, 16 Jul 4, 18 150 ... 34 0.003 -85 ... M1.8 1 1 360 6622012 Jul 5, 21 Jul 5, 22 161 ... 46 BG -73 0.0002 M1.6 1 2 94 9802012 Jul 6, 23 Jul 7, 01 165 ... 50 0.3 -69 0.004 X1.1 1 2 360 18282012 Jul 8, 16 Jul 8, 16 -171 ... 74 0.1 -46 0.1 M6.9 1 2 157 14952012 Jul 12, 15 Jul 12, 17 109 0.2 -06 0.8 -126 BG X1.4 1 2 360 8852012 Jul 13, 19 Jul 13, 20 -115 BG 130 BG 10 0.004 ... 1 0 115 3862012 Jul 17, 12 Jul 17, 15 180 ... 65 0.8 -55 0.001 M1.7 1 2 176 9582012 Jul 18, 06 Jul 18, 06 -75 0.001 170 BG 50 0.1 ... 1 1 360 8732012 Jul 19, 04 Jul 19, 08 -145 ... 100 0.8 -20 BG M7.7 1 2 360 16312012 Jul 23, 02 Jul 23, 03 -105 0.5 140 0.15 20 50 ... 1 2 360 20032012 Aug 10, 09 Aug 10, 13 -97 ... 147 0.0012 25 0.008 ... 1 0 251 4642012 Aug 19, 04 Aug 19, 05 -71 ... 173 0.0002 50 0.08 ... 1 0 224 5772012 Aug 19, 18 Aug 19, 18 -71 0.001 173 0.0002 50 0.4 ... 1 0 360 6122012 Aug 21 20 Aug 21, 21 -56 0.0008 -172 ... 65 BG ... 1 2 360 521


Multi-sp

acecra

ftSE

Pev

ents



2012 Aug 31, 19 Aug 31, 20 74 20 -42 0.037 -165 0.0008 C8.4 1 2 360 14422012 Sep 8, 10 Sep 8, 11 -99 ... 145 0.014 20 0.002 ... 1 1 360 7342012 Sep 19, 11 Sep 19, 12 -51 ... -168 ... 67 0.002 ... 1 1 360 6162012 Sep 19, 22 Sep 19, 22 -46 ... -163 ... 72 0.04 ... 1 0 141 4962012 Sep 20, 15 Sep 20, 15 -40 0.2 -158 0.003 77 5 ... 1 2 360 12022012 Sep 27, 10 Sep 27, 10 -84 ... 158 BG 33 0.15 ... 1 2 360 13192012 Sep 27, 23 Sep 28, 00 152 0.3 34 0.2 -91 BG C3.7 1 2 360 7552012 Oct 7, 06 Oct 7, 12 124 ... 5 0.0002 -121 ... B4.5 1 0 149 6632012 Oct 14, 00 Oct 14, 01 -18 0.008 -138 0.0002 96 0.06 ... 1 1 360 9872012 Nov 8, 02 Nov 8, 03 35 0.006 -88 ... 145 0.0003 M1.7 1 0 360 8552012 Nov 8, 11 Nov 8, 11 -68 BG 168 0.015 41 0.4 ... 1 0 360 9722012 Nov 16, 06 Nov 17, <19 10 0.005 -114 ... 118 ... C1.4 1 0 360 7752012 Nov 20, 12 Nov 20, 13 -135 ... 10 0.0001 -118 ... M1.7 1 0 360 6192012 Nov 21, 15 Nov 21, 17 124 0.001 -1 0.0003 -128 ... M3.5 0 0 360 5292012 Nov 23, 23 Nov 23, 23 -102 ... 132 0.0003 5 0.03 ... 1 2 360 11862012 Nov 24, 13 Nov 24, 14 154 ... 28 0.0003 -100 ... C3.3 1 0 84 2372012 Dec 2, 15 Dec 2, 16 -56 ... 177 0.0005 50 0.005 ... 1 0 360 6782012 Dec 5, 00 Dec 5, 03 52 0.003 -75 ... 157 0.0002 C1.7 1 1 231 9632012 Dec 14, 01 Dec 14, 03 169 ... 40 0.01 -88 ... ... 1 0 149 7632013 Jan 6, 22 Jan 6, 23 -40 ... -173 ... 58 0.008 ... 1 0 211 6672013 Jan 16, 18 Jan 16, 20 -149 ... 77 0.005 -52 ... C2.2 1 0 250 6482013 Feb 6, 00 Feb 6, 03 118 0.002 -19 0.0001 -149 0.0004 C8.7 1 0 271 18672013 Feb 26, 09 Feb 26, 13 -89 ... 131 0.01 0 0.002 ... 1 1 360 9872013 Mar 5, 03 Mar 5, 03 -1 0.4 -141 0.006 88 30 ... 1 1 360 13162013 Mar 15, 05 Mar 15, 07 129 0.006 -12 0.004 -144 ... M1.1 1 2 360 10632013 Apr 11, 06 Apr 11, 07 130 5 -12 2 -146 0.001 M6.5 1 2 360 8612013 Apr 21, 06 Apr 21, 08 -94 ... 124 0.02 -10 0.001 ... 1 0 360 9192013 Apr 24, 21 Apr 24, 22 -44 0.001 175 0.01 40 0.1 ... 1 0 360 5942013 Apr 28, 20 Apr 28, 22 179 ... 37 0.0004 -97 ... C4.4 1 0 91 4972013 May 1, 02 May 1, 04 37 0.009 -105 ... 120 0.007 ... 1 1 360 7622013 May 2, 04 May 2, 06 168 BG 26 0.0005 -109 BG M1.1 1 0 99 6712013 May 3, 17 May 3, 20 61 0.003 -81 ... 144 0.002 M5.7 0 0 274 8582013 May 10, 18 May 10, 20 -78 ... 140 0.0002 5 0.006 ... 1 0 120 6172013 May 13, 01 May 13, 02 39 0.2 -103 BG 121 ... X1.7 1 1 360 12702013 May 13, 15 May 13, 16 47 1.0 -95 0.01 129 0.004 X2.8 1 2 360 1850


I.G

.R

ichard

son

etal.



2013 May 15, 01 May 15, 06 78 BG -64 0.3 160 BG X1.2 1 2 360 13662013 May 22, 12 May 22, 13 -149 0.003 70 20 -67 0.06 M5.0 1 2 360 14662013 Jun 17, 02 Jun 17, 03 81 0.002 -59 ... 161 ... C1.2 0 ... ... ...2013 Jun 21, 02 Jun 21, 03 67 0.6 -73 0.06 147 0.001 M2.9 1 ... 160* 1249*2013 Jul 1, 20 Jul 1, 20 50 0.015 -90 ... 130 0.004 ... 1 ... 188* 542*2013 Jul 3, 07 Jul 3, 07 58 0.002 -82 ... 137 ... M1.5 1 ... 102* 580*2013 Jul 22, 06 Jul 22, 06 -50 0.001 172 0.001 29 0.04 ... 1 ... 360 M 1032 M2013 Aug 17, 19 Aug 17, 19 168 ... 30 0.002 -174 ... M3.3 -1 ... 156* 726*2013 Aug 19, 23 Aug 20, 00 -48 0.4 174 0.01 30 3 ... 1 ... 144* 637*2013 Aug 30, 02 Aug 30, 02 95 0.03 -43 0.0003 172 ... C8.2 1 ... 360 M * 1071 M*2013 Sep 24, 19 Sep 25, 00 79 0.0002 -60 0.0001 153 ... ... 1 ... 126* 661*2013 Sep 29, 21 Sep 29, 22 164 0.02 25 0.5 -122 0.003 C1.2 1 ... 360 M 897 M2013 Oct 5, 08 Oct 5, 08 16 0.02 -123 ... 90 0.5 ... 1 ... 324* 640*2013 Oct 11, 07 Oct 11, 07 44 0.3 -96 0.002 117 5 M1.5 1 ... 360 M 1251 M2013 Oct 22, 21 Oct 22, 22 141 0.002 00 0.01 -148 ... M4.2 1 ... 360 M 658 M2013 Oct 25, 08 Oct 25, 08 69 0.3 -73 0.001 139 ... X1.7 1 ... 360 M 715 M2013 Oct 25, 14 Oct 25, 15 74 1.0 -68 0.008 144 0.002 X2.1 1 ... 360 M 1476 M2013 Oct 28, 04 Oct 28, 04 -147 BG 71 0.04 -77 0.002 M5.1 1 ... 50* 1180*2013 Oct 28, 15 Oct 28, 16 114 0.2 -28 BG -176 BG M4.4 1 ... 132* 589*2013 Nov 2, 04 Nov 2, 04 -90 0.04 127 0.02 -21 1.5 ... 1 ... 360* 589*2013 Nov 4, 05 Nov 4, 06 -49 BG 168 BG 20 1.2 ... 1 ... 134* 806*2013 Nov 6, 23 Nov 7, 01 -118 BG 99 0.01 -50 BG M1.8 1 ... 360 M 1358 M2013 Nov 7, 10 Nov 7, 10 0 12 -143 BG 68 1.3 ... 1 ... 360 M 1820 M2013 Nov 19, 10 Nov 19, 11 -146 0.001 69 0.04 -80 0.0006 X1.0 1 ... 360 M 959 M2013 Nov 26, 15 Nov 26, 15 -44 ... 169 ... 20 0.0006 ... 1 ... ... ...2013 Nov 30, 05 Nov 30, 06 -63 ... 150 0.0009 0 0.0005 ... 1 ... 98* 371*2013 Nov 30, 15 Nov 30, 15 0 0.0002 -146 BG 63 BG ... 0 ... 140* 496*2013 Dec 5, 10 Dec 5, 14 40 0.003 -108 BG 102 0.001 ... 0 ... 132* 422*2013 Dec 7, 07 Dec 7, 09 -164 ... 49 0.0004 -100 ... M1.2 1 ... 132* 919*2013 Dec 9, 19 Dec 9, 23 78 0.0005 -70 ... 139 0.0003 C1.9 0 ... 166* 606*2013 Dec 12, 03 Dec 12, 04 -165 ... 46 0.001 -104 ... C4.6 1 ... 146* 672*


Multi-sp

acecra

ftSE

Pev

ents



2013 Dec 13, 20 Dec 13, 20 0 ... -149 ... 61 0.001 ... 1 ... 144* 456*2013 Dec 14, 02 Dec 14, 02 3 ... -146 ... 64 0.006 ... 1 ... ... ...2013 Dec 14, 06 Dec 14, 06 5 ... -144 0.002 66 0.02 ... 1 ... 72* 644*2013 Dec 16, 21 Dec 16, 22 -101 ... 110 0.0002 -40 ... C3.1 1 ... 102* 508*2013 Dec 26, 03 Dec 26, 04 -10 0.8 -161 0.03 49 0.8 ... 1 ... 202* 856*2013 Dec 28, 17 Dec 28, 18 -78 BG 130 0.3 -20 BG ... 1 ... 218* 1420*

Column 1: Date, hour of solar event; Column 2: Date, hour of SEP onset at spacecraft withearliest onset; Column 3: Longitude of solar event relative to STEREO B where positive(negative) indicates that the event is west (east) of the spacecraft; Column 4: ∼ 25 MeVproton intensity ((MeV s cm2 sr)−1) at STEREO B. ‘...’ indicates that the SEP event wasnot detected above ∼ 10−4 (MeV s cm2 sr)−1, and ‘BG’ that a high background from aprevious event may have obscured the onset; Column 5: Longitude of solar event relative toEarth/SOHO; Column 6: ∼ 25 MeV proton intensity at Earth/SOHO; Column 7: Longitudeof solar event relative to STEREO A; Column 8: ∼ 25 MeV proton intensity at STEREO A;Column 9: GOES soft X-ray flare intensity; Column 10: Type III radio emissions observed byWIND/WAVES or STEREO A/B SWAVES (0 = no; 1 = yes); Column 11: Type II emissionsobserved by WIND/WAVES or STEREO A/B SWAVES (0 = no; 1 = yes; 2 = “IP” type IIextending below 1 MHz; available to May 2013 at the time of writing); Column 12: CMEwidth, and Column 13: CME speed. The CME parameters are from the CUA/GSFC/CDAWLASCO catalog (available to May 2013 at the time of writing) except where indicated by ‘*’= from the CACTUS LASCO catalog, ‘M’ = from the real-time LASCO halo CME reports,‘A’ = from the CACTUS STEREO A catalog, and ‘B’ = from the CACTUS STEREO Bcatalog.



STEREO SWAVES. As discussed by Cane and Erickson (2005), many of thetype II radio events in this list are minor, do not last for very long and cover alimited frequency range, indicating that the associated shocks are not particu-larly strong, whereas strong interplanetary shocks typically produce broad bandradio emissions (“IP type II events”; Cane et al., 1982; Cane and Stone, 1984;Cane, 1985) starting below a few MHz. Thus, if the type II emissions extendedbelow 1 MHz, indicative of an IP type II event, this is denoted by ‘2’ rather thanby ‘1’. At the time of writing, the type II event list is only available up to May2013.

The next two columns give the angular width and speed of the associatedCME. For consistency with our previous studies (e.g., Cane, Richardson andvon Rosenvinge (2010a)), we show CME widths and speeds from the CatholicUniversity of America/Goddard Space Flight Center CDAW CME catalog (http://cdaw.gsfc.nasa.gov/CME list/; Yashiro et al., 2004) compiled from observa-tions made by the LASCO coronagraphs on SOHO, while recognizing thatSOHO may not be the best spacecraft (e.g., closest to quadrature) to observethe CME speed and width with minimal plane of the sky projection. Sincethese values are only available until May 2013 at the time of writing, valuesfrom the CACTUS LASCO CME database (http://sidc.oma.be/cactus/), indi-cated by an asterisk, or from the SOHO real-time observer e-mail halo CMEalerts (http://umbra.nascom.nasa.gov/lasco/observations/halo/soho-halo-alerts/),indicated by ‘M’, are included for more recent events. In a few cases where thereare no LASCO observations, CACTUS CME widths and speeds from STEREOA or B are listed, indicated by ‘A’ or ‘B’, respectively. Data gaps are indicatedby “DG”.

2.2. Example Events

Tables 1–7 indicate that individual SEP events were observed by just one space-craft, or by two or three. (While the number of observing spacecraft does ofcourse depend on the spacecraft locations relative to the solar event as wellas on the longitudinal extent of the SEP event, we will show below that thisparameter does order the SEP properties to some degree.) Figure 4 shows two ofthe most intense probable single-spacecraft proton events, on 9 and 11 May 2011.The first was associated with a solar event at ∼ 21 UT on 9 May near centralmeridian with respect to STEREO B, which observed the particle enhancement,and at E94o (i.e. just behind the east limb) as viewed from Earth. The locationsof STEREO A and B and the Earth with respect to the solar event are shownbelow the data panel. Nominal Parker spiral field lines are drawn to the Sunusing the observed solar wind speed at each location. A C5.4 X-ray flare and afast (1318 km s−1) and wide (292o) CME (values from LASCO) were observed,together with IP type II and type III radio emissions. The 14–24 MeV protonintensities at STEREO A (for which the solar event was at W174o) and SOHOsuggest that this event may just have been detected at these spacecraft, butat ∼ 25 MeV, the intensity did not increase above our threshold of ∼ 10−4

(MeV s cm2 s)−1.The second SEP event in Figure 4, detected by ERNE, was associated with

a solar event at ∼ 02 UT on 11 May at W53o with respect to Earth, which was



10-5

10-4

10-3

10-2

10-1

100STEREO A 14-24 MeV Protons 24-41 MeV Protons

10-5

10-4

10-3

10-2

10-1

Inte

nsity (

MeV

s c

m**

2 s

r)**

-1

10-5

10-4

10-3

10-2

10-1

2011

129

9

130

10

131

11

132

12

133

13

May

DOY

�

STEREO B 14-24 MeV Protons 24-41 MeV Protons

SOHO/ERNE 14-24 MeV Protons 24-41 MeV Protons

1.0 AU

STEREO ASTEREO B

Earth/SOHO 1.0 AU

STEREO ASTEREO B

Earth/SOHO

Figure 4. Intensities of protons at 14 – 24 MeV (black) and 24–41 MeV (red) observed bythe HET instruments on STEREO A and B (top and bottom data panels, respectively) andthe ERNE instrument on SOHO (middle panel), for two “one-spacecraft” events in May 2011from two different active regions. The locations of the spacecraft relative to the solar events areshown below. Nominal Parker spiral magnetic field lines passing each spacecraft are indicated.In both cases, the particle event is observed at the spacecraft that is best connected to thesolar event.

nominally well-connected to the solar event by the IMF. The event was at E33o

with respect to STEREO A, which showed no significant increase of ∼ 25 MeVprotons, and at W147o with respect to STEREO B, where it is possible thatthe decay of the preceding event might have obscured a weak enhancementfrom this event. A B8.1 X-ray flare and a 745 km s−1 LASCO CME with awidth of 225o were associated with the particle event, as well as type III radioemissions (occulted at STEREO B as might be expected), but no reported typeII. Thus, both these single-spacecraft particle events were associated with wideand reasonably fast (in one case, exceeding 1000 km s−1) CMEs, and the particleevent was only detected by the spacecraft that was best connected to the solarevent.

Figure 5 shows an example of a three-spacecraft SEP event that occurred earlyon 21 March 2011. The solar event was observed by STEREO A at approximately



10-410-2100102104106

STEREO A 14-24 MeV Protons 24-41 MeV Protons

10-410-2100102104

Inte

nsity (

MeV

s c

m**

2 s

r)**

-1

10-410-2100102104

2011

79

20

80

21

81

22

82

23

83

24

Mar

DOY

�



14*(0.7-4 MeV e -0.015)

14*(0.7-4 MeV e -0.02)

0.67-3.0 MeV e -0.006

1.0 AU

STEREO ASTEREO B

Earth/SOHO

Figure 5. A “three-spacecraft” event on 21 March 2011. Intensities of 14 – 24 and 24–41 MeVprotons and 0.7–4 MeV electrons (corrected for background and the HET calibration error)observed by the STEREO A and B HETs are shown in the top and bottom panels, respectively.The middle panel shows proton intensities at the same energies from the ERNE instrumenton SOHO and the 0.67–3 MeV electron intensity from EPHIN. The particle event rose mostpromptly, and was most intense, at STEREO A which was best connected to the solar event,whereas a delayed rise and weak event, barely above background, was observed at STEREO B,which was connected ∼ 170o from the solar event. Though the event was on the far side ofthe Sun (W138o) relative to Earth, a clear, reasonably prompt event onset was observed byERNE. Particle intensities were around two orders of magnitude lower than at STEREO A.The right-hand panel shows 2.5 kHz–16 MHz radio observations from STEREO A (top) andB (bottom) in a “mirrored” format, illustrating the bright, long-duration type III emissionsthat accompanied this event. The occultation at high frequencies at STEREO B (and alsoWIND, not shown) by the limb of the Sun is consistent with the spacecraft locations withrespect to the solar event. Slower drifting Type II emission is also evident at STEREO A at02:30–04:30 UT.

W50o with respect to this spacecraft, which was well-connected to the eventand observed a prompt increase in particle intensity. The HET observed the0.7–4 MeV electron onset at 02:33 UT ±1 minute. (HET electron intensitiesin this and other event figures are corrected as discussed above in relationto Figures 3(f) and (g).) The electron onset occurred ∼ 13 minutes after theonset of bright, long duration, type III radio emissions observed by STEREO ASWAVES at 02:20 UT; see the right-hand panel in Figure 5. The emissions wereocculted at high frequencies by the limb of the Sun when observed at STEREOB (and also at WIND, not shown), consistent with the configuration of thespacecraft relative to the solar event shown in Figure 5. Interplanetary Type IIradio emission was also observed by STEREO A at 02:30–04:30 UT. HET Aobserved the 24–41 MeV proton intensity increase from 03:14 UT ±2 minutes,i.e. ∼ 40 minutes later than the near-relativistic electron onset. A 25 MeV



(40 MeV) proton would take ∼ 45 (35) minutes to travel from the Sun alonga nominal 1.2 AU spiral magnetic field line, suggesting that this delay maybe largely accounted for by the proton propagation time from the Sun. Thesolar event was at W138o relative to Earth. Nevertheless, a reasonably promptparticle increase was also observed (0.67–3 MeV electron and 24–41 MeV protononsets were at 03:14 UT ±5 minutes and 03:40 UT ±5 minutes, respectively).However, the particle intensities were around two orders of magnitude lower thanat STEREO A. The event was at E127o with respect to STEREO B, which wasconnected to the Sun ∼ 170o from the event. HET B observed a weak, slowlyrising increase that we attribute to this event. Proton intensities were another∼two orders of magnitude lower than at the Earth. Thus, the particle event wasmost prompt and most intense at the best-connected spacecraft, and the onsetdelay increased and the intensity decreased as the spacecraft connection to theevent became poorer. The associated CME had a mean speed of 833 km s−1

and width of 260o based on CACTUS analysis of observations from STEREO B,the spacecraft closest to quadrature, while the LASCO CDAW catalog classifiesthis as a full halo CME with a speed of 1341 km s−1. See Rouillard et al. (2012)for a detailed study of the relationship between the solar eruption, CME andenergetic particle onset of this event.

The three-spacecraft event late on 3 November 2011 (Figure 6) was remarkablein that the proton and electron intensities showed relatively prompt increases atall three spacecraft despite their wide separation. STEREO B observations showthat the associated solar event was at ∼ 50o east of this spacecraft. The particleevent was again most intense at the best-connected spacecraft (STEREO A) andwas around two orders of magnitude weaker at the other spacecraft. Figure 7shows an expanded view of the beginning of the event using 1 minute averageddata from the STEREO HETs and 5 minute averaged ERNE and EPHIN ob-servations. The electron onset occurred at STEREO A at 22:33 UT, around 13minutes after the beginning of the associated type III emissions (interplanetarytype II emissions were also observed by STEREOs A and B starting at 22:35 UT).The 24–41 MeV proton intensity at STEREO A increased around 52 minuteslater. The proton intensity increased at the Earth around 25 minutes later, andat a similar time at STEREO B. Thus, within 25 minutes after protons weredetected at the best-connected spacecraft, they were also detected at the othertwo, poorly–connected, spacecraft. Dividing the angular distance between thefoot points of the field lines to STEREO A and SOHO by 25 minutes suggestsan angular “propagation speed” of 5.3o/minute, corresponding to a speed of 1070km s−1 at the surface of the Sun. This turns out to be reasonably comparableto the expansion speed of the associated LASCO CME (991 km s−1) givenin the CUA/GSFC catalog though this may be fortuitous (see the discussion ofFigure 17 below). This was also a full halo CME. CACTUS CME speed estimatesare 625 km s−1 (LASCO), 781 km s−1 (STEREO A, closest to quadrature)or 694 km s−1 (STEREO B). We note that despite the widespread, relativelyprompt onset, this particle event was not associated with an especially fastCME. The CME speed is, for instance, comparable to the speeds of the CMEsassociated with the single-spacecraft events in Figure 4. We have also examinedthe possibility that two or more solar events were involved in this widespread



10-410-2100102104106


10-410-2100102104

Inte

nsity (

Me

V s

cm

**2

sr)

**-1

10-410-2100102104

2011

307

3

308

4

309

5

310

6

311

7

Nov

DOY

�



14*(0.7-4 MeV e -0.015)

14*(0.7-4 MeV e -0.02)

0.67-3.0 MeV e -0.006

STEREO ASTEREO B

Earth/SOHO

Figure 6. A three-spacecraft event on 3 November 2011 in which the particle intensity rosepromptly at all three spacecraft, including at the Earth for which the solar event was at E152o.Again note that the event was most intense at the best-connected spacecraft, STEREO A.

particle event, such as “sympathetic flares” (Richardson, 1936; Moon et al., 2002;Schrijver and Title, 2011; Schrijver et al., 2013) triggered by the initial event.However, solar imaging, WAVES/SWAVES radio observations, X-rays and otherobservations show no evidence that multiple events were involved. In particular,an X1.9 flare at E63o relative to the Earth with peak intensity at 20:27 UT,around two hours earlier, played no role since STEREO B was well-connectedto this flare and did not detect any particle increase. In addition, there were notype II or type III radio emissions associated with this flare, and the STEREOB COR 1 coronagraph only observed a narrow (∼ 30o width) CME that doesnot appear in the CDAW or CACTUS catalogs.

Particle increases associated with closely-timed solar events do occur occasion-ally, and consideration of the solar, radio and particle data together can typicallylead to a likely interpretation. For example, two particle onsets from different



10-510-410-310-210-1100101102103

STEREO A 14-24 MeV H 24-41 MeV H 0.7-4 MeV e-

10-510-410-310-210-1100101

Inte

nsity (

Me

V s

cm

**2

sr)

**-1

10-510-410-310-210-1100101

2011

21 22 23 0

308

4

1 2

Nov

DOY

UT.

STEREO B 14-24 MeV H 24-41 MeV H 0.7-4 MeV e-

SOHO/ERNE 14-24 MeV H 24-41 MeV H EPHIN 0.67-3 MeV e-

Figure 7. Expanded view of the onset of the event on 3 November 2011, showing 1 min-ute-averaged data from the STEREO HETs, and 5 minute-averaged data from SOHO ERNEand EPHIN. The vertical line indicates the onset time of type III radio emissions observed bySTEREO A SWAVES. Electrons were first detected at STEREO A around 13 minutes afterthis time, followed by 24–41 MeV protons 52 minutes later. Similar energy protons arrived atSTEREO B and Earth with an additional delay of ∼ 25 minutes.

active regions occurred within 7 hours on 7 March 2011 (Figure 8). The first wasassociated with an M1.9 flare at E21o relative to the Earth, accompanied by a698 km s−1 partial halo CME (LASCO/CUA) and was most intense at STEREOB, which was well-connected to the event. A weak increase was seen at Earth,while there is no clear evidence of one at STEREO A. Type III radio emissionscommenced at 14:15 UT, while type II emission was observed at STEREO Band WIND from 14:30 UT. A second solar event was associated with an M3.7flare located at W53o relative to Earth and a fast (2125 km s−1) halo CMEthat was best connected to the Earth, where SOHO observed a large promptonset. STEREO A, for which the event was at E35o, observed a more slowlyrising particle enhancement, peaking at a shock that passed the spacecraft on9 March, that was probably associated with the same solar event/CME. Sucha profile is typical of an eastern event (cf., Cane, Reames and von Rosenvinge,1988). Electron observations suggest that STEREO B also detected particlesfrom this event, and they probably also contributed to the proton enhancement.However, the intensity profile at STEREO B is complicated by the passage of ashock, sheath and interplanetary coronal mass ejection (ICME) on 7–8 March.In particular, the brief intensity decrease on 8 March is associated with passageof the ICME (Leske et al. (2013) reported field-aligned bidirectional ∼ 5 MeV



10-510-410-310-210-1100101102103


10-510-410-310-210-1100101

Inte

nsity (

MeV

s c

m**

2 s

r)**

-1

10-510-410-310-210-1100101

2011

0

66

7

8 16 0

67

8

8 16 0

68

9

8 16 0

69

10

Mar

DOY

UT.

ICME

˛�



14*(0.7-4 MeV e -0.015)

14*(0.7-4 MeV e -0.02)

0.67-3.0 MeV e -0.006

STEREO A

STEREO B

Earth/SOHO

STEREO A

STEREO B

Earth/SOHO

Figure 8. Observations for two particle events from different active regions on 7 March 2011.Vertical lines give the onset times of type III radio emissions. The first event, from E21o

(relative to the Earth), was most prominent at STEREO B, the best-connected spacecraft.The second, from W53o, was best connected to the Earth, but was also observed at the otherspacecraft. At STEREO A, the intensity peaks at passage of the shock on 9 March that wasprobably associated with this event.

proton flows, characteristic of ICMEs (e.g., Zurbuchen and Richardson (2006)and references therein) in this region). This second particle event was associatedwith bright, long duration, type III radio emissions (occulted at STEREO B)commencing at 20:00 UT and IP type II emission from 20:00 UT observed bySTEREO A and WIND.

Our final example, which occurred one year after the previous event, is acase where two CMEs erupted in the vicinity of AR 1429 separated by ∼ 1hour. On 7 March 2012, an X5.4 flare (onset: 00:02 UT; peak: 00:24 UT; end:00:40 UT) at E27o with 3B H-α emission was followed by an X1.3 flare (onset:01:05 UT; peak: 01:14 UT; end: 01:23 UT) at E17o. Two fast halo LASCOCMEs were associated with these events, first observed above the occulting disk



10-310-210-1100101102103

STEREO A 14-24 MeV H 24-41 MeV H 0.7-4 MeV e-

10-310-210-1100101102

Inte

nsity (

MeV

s c

m**

2 s

r)**

-1

10-310-210-1100101102

2012

23 0

67

7

1 2 3 4 5 6 7

Mar

DOY

UT.

STEREO B 14-24 MeV H 24-41 MeV H 0.7-4 MeV e-

SOHO/ERNE 14-24 MeV H 24-41 MeV H EPHIN 0.67-3 MeV e-

STEREO A

STEREO B

Earth/SOHO

Figure 9. Observations of the onset of the 7 March 2012 particle event. Shaded regionsindicate the durations of two X-class X-ray flares at E27o (also evident as contamination inthe EPHIN electron channel) and E17o relative to Earth, both accompaniedby fast halo CMEs.The vertical line gives the onset time of the bright type III emissions observed by STEREOsA and B (right panel) and at the Earth which accompanied the first event. Observations atSTEREO B (1 minute averages) clearly demonstrate that the onset of particle accelerationwas associated with the first event rather than the second; there is no additional feature in theintensity–time profile that might be related to the second event.

at 00:24 UT (with a speed of 2684 km s−1), and at 01:30 UT (1825 km s−1).Figure 9 shows 1 minute-averaged observations from the STEREO HETs and 5minute-averaged SOHO ERNE and EPHIN data during the onset of this event.The electron intensity at STEREO B, the best-connected spacecraft, was clearlyincreasing by ∼ 00 : 45 UT. As is typical (cf., the events described above,and further discussion below), proton onset was delayed by ∼ 40 minutes, to∼ 01 : 25 UT. These onset times are clearly consistent with an association withthe first solar event (shaded regions indicate the X-ray flare intervals for the twoevents) and are inconsistent with an association with the second solar event.While the particle event was eventually observed at the other spacecraft, notethat it is the well-connected STEREO B observations that provide unambiguousevidence of the correct solar association. We also note that only the first eventwas accompanied by bright type III radio emissions, again consistent with thisbeing the site of particle acceleration and release. Interplanetary type II emissioncommenced at 01:00 UT, also ahead of the second event.

The occurrence of these two closely-spaced CMEs might be expected to be anideal situation for particle acceleration by multiple shocks (e.g., Pomerantz andDuggal, 1974; Levy, Duggal and Pomerantz, 1976; Cliver and Kallenrode, 2001)or multiple CMEs (e.g., Gopalswamy et al., 2002; Gopalswamy et al., 2004; Li



Cycle 23 Cycle 24

10-3

10-1

101

103

105

~2

5 M

eV

Pro

ton

In

t.

0

3

6

9

No

. E

ve

nts

/ro

t

0

40

80

120

160

SS

N

010203040506070

2000 2005 2010

No

. E

ve

nts

/ye

ar Events/year < 10-1

Events/year > 10-1Total events/yearSTA, STB, Earth

Figure 10. Summary of the rate of > 25 MeV proton events at Earth during January1996–December 2013. The top panel shows the event peak intensity at ∼ 25 MeV (in(MeV s cm2 sr)−1) while the second panel gives the number of events/Carrington rotationwith intensities >∼ 2 × 10−4 (MeV s cm2 sr)−1 and the three-rotation running mean (redgraph). The monthly sunspot number is shown in panel three. Panel four gives the yearlynumber of events at Earth (green) and those with intensities above (red) or below (black)0.1 (MeV s cm2 sr)−1. The blue graph shows the number of individual events identified atSTEREO A or B, or at the Earth.

et al., 2012) to take place. We note that Cliver and Kallenrode (2001) concluded

that acceleration by multiple shocks does not necessarily require the shocks to be

converging, so the fact that the first CME was faster than the second according

to the CDAW LASCO CME database would not necessarily rule out this process

from occurring. It has also been suggested that a preceding CME may provide a

“seed” population for acceleration by the shock of a second CME or may in some

way “precondition” the solar wind through which the second CME shock travels

to promote particle acceleration. However, the observations of this particle event

onset clearly indicate that particle acceleration was associated with the first solar

event/CME and that the second CME was not involved. There is also no obvious

indication in the STEREO B intensity–time profile of a second particle injection

associated with the second solar event. This, together with the absence of type

III emissions (as discussed above and also below, the vast majority of > 25 MeV

proton events are accompanied by such emissions) also suggest that the second

solar event played no role in particle acceleration near the Sun.



3. Summary of SEP Events

3.1. Occurrence rate

Figure 10 summarizes the occurrence of > 25 MeV proton events at Earth from1996 to December 2013, updated from Cane, Richardson and von Rosenvinge(2010a) and Cane, Richardson and von Rosenvinge (2010b). The top panel showsthe event peak intensity at ∼ 25 MeV. The second panel gives the numberof SEP events/Carrington rotation and the three-rotation running mean (red),while the third panel shows the monthly sunspot number. The SEP event rateclearly tends to follow solar activity levels through Cycle 23 (though with SEPevents persisting until well into the declining phase), the subsequent extendedsolar minimum, and into Cycle 24. The sunspot number in Cycle 24 reacheda peak of 97 in November 2011 dominated by northern hemisphere sunspots(Chowdhury, Choudhary, and Gosain, 2013), well below the maximum in Cycle23, and then declined to around 60 for much of 2012–2013. Consistent with thedecrease in the level of solar activity, the SEP rate declined from around mid-2012 into 2013, though fairly intense events were still observed occasionally, inparticular in April–May 2013. Towards the end of 2013, the sunspot number (nowdominated by southern hemisphere sunspots) appears to increasing once again,and a corresponding increase in the SEP rate is also evident. We therefore suggestthat the decline in activity in 2012–2013 is an example of the temporary decreasein energetic solar activity, including SEP events, often termed the “GnevyshevGap”, that is frequently found near solar cycle maxima (e.g., Gnevyshev, 1967;Gnevyshev, 1977; Feminella and Storini, 1997; Storini et al., 2003; Bazilevskaya,Makhmutov, and Sladkova, 2006; Norton and Gallagher, 2010; Richardson andCane, 2012, and references therein).

An interesting difference between Cycles 23 and 24 is that 25 MeV protonevents were present in Cycle 23 from around two months after the time of sunspotminimum in May 1996 whereas the first event at Earth in Cycle 24 was detectedwith a delay of a year after sunspot minimum, which occurred in December 2008.The SEP occurrence rate (second panel) then rose somewhat more rapidly thanduring the rise of Cycle 23. The yearly SEP event rates at the Earth (bottompanel) appear to have been reasonably comparable in the two cycles based onthe observations so far, being only ∼ 20% below the highest rates in Cycle 23.However, the yearly rates for 2013 indicate a decline from the previous year. Theblue graph in the bottom panel shows the number of individual events (as inTables 1–7) observed at STEREO A or B or at the Earth. Around 30% of theseevents were not detected at the Earth. Again there is an indication of a fall inthe SEP event rate in 2013 by around a third from the rate in 2012.

3.2. Solar event properties

When discussing the properties of SEP events at the STEREO spacecraft and atEarth, and the associated solar events, we will generally focus on the subset ofevents detected from December 2009 when the STEREO spacecraft were ≥ 65o

from Earth and the SEP rate increased in Cycle 24, to December 2012, when the



Figure 11. The left-hand figure shows the distribution of > 25 MeV proton events detectedby one, two or three spacecraft, and those for which the number of detecting spacecraft isunclear, from December 2009, when the STEREO spacecraft were ∼ 65o from the Sun-Earthline, to December 2012, when the STEREO spacecraft were separated by 99o. The right-handfigure shows the percentages of events that originated behind the east limb, on the front sideof the Sun, or behind the west limb, relative to STEREO B, the Earth, or STEREO A, andfor the events in 1997–2006 discussed by Cane, Richardson and von Rosenvinge (2010a).

STEREO spacecraft were separated by 99o on the far side of the Sun. Althoughthe spacecraft constellation was changing during this period, the spacecraft werewell separated, with the STEREO spacecraft positioned approximately abovethe limbs of the Sun as observed from Earth. The left-hand panel of Figure 11shows the percentages of the ∼ 25 MeV proton events during this interval thatwere detected by only one, only two or all three spacecraft For 13% of theevents, the number of observing spacecraft is unclear because of data gaps orhigh background from preceding events. Otherwise, 36% of the SEP events wereobserved by only one spacecraft, 34% at only two spacecraft, and 17% at all threespacecraft. (Considering all the events in Tables 1–7, 36% were observed by onespacecraft, 34% by two spacecraft, and 21% by all three.) As will be discussedbelow, the number of spacecraft that detected an SEP event, which gives acrude indication of the longitudinal extent of the event but is appropriate forthis study focusing on identifying events at the different spacecraft, does tend toorganize the event properties. The right-hand panel summarizes the percentageof the events that originated behind the east or west limbs, or on the front sideof the Sun from the viewpoint of STEREO A, B or the Earth. Prior to theSTEREO spacecraft approaching the limbs of the Sun, allowing observations ofthe far side of the Sun, the detection at Earth of SEP events originating behindthe limbs of the Sun could only be inferred indirectly, for example from CMEsoriginating behind the limb, occulted type III radio emissions, and the presenceof major active regions that had previously rotated over the west limb, or lateremerged over the east limb. Thus, this panel also shows similar results basedon the events detected during 1997–2006 at Earth in Cane, Richardson and vonRosenvinge (2010a). The distribution of events is similar in each case, with ∼ 60–75% originating on the front side, ∼ 20–30% behind the west limb, and ∼ 4–8%behind the east limb. The east-west asymmetry reflects the favored connectionto the western hemisphere of the Sun by the spiral interplanetary magnetic field.

Figure 12 shows the ∼ 25 MeV proton intensity (including observations fromall three spacecraft) plotted against the longitude of the solar event with respectto the observing spacecraft for events from December 2009 to, in this case, De-cember 2013, providing a summary of all but the few earliest events in the study



-100 0 100

10-3

10-2

10-1

100

101

25

Me

V P

roto

n I

nte

nsity

Flare Longitude wrt. S/C(deg.)

No. of Spacecraft

1

2

3

Figure 12. Peak intensities of ∼ 25 MeV proton events (in (MeV s cm2 sr)−1) plotted againstthe longitude of the solar event relative to the observing spacecraft, for events in December2009–December 2013, when the STEREO spacecraft were separated by > 58o in longitude.The symbol type indicates whether the event was detected at one (green cross), two (red circle)or three (blue square) spacecraft (a point is plotted for each observing spacecraft for multi-ple–spacecraft events). Three-spacecraft events with intensities above 10−2 (MeV s cm2 sr)−1

from behind the east limb are circled. The corresponding intensities at the best-connectedspacecraft are similarly indicated.

period. The symbol/color indicates the number of spacecraft that observed eachevent; 325 observations are included in the figure. One-spacecraft events (greencrosses) predominantly originate on the western hemisphere, and typically havepeak intensities < 10−2 (MeV s cm2 sr)−1 . Thus, they are typically weaker eventsthat are detected when a spacecraft is favorably connected to the event. Two-spacecraft events (red circles) are around two orders of magnitude more intensethan single-spacecraft events at well-connected longitudes, and may originatefrom behind the west limb to the eastern front-side hemisphere. Three-spacecraftevents typically have intensities around 1–2 orders of magnitude larger than two-spacecraft events and form the majority of events that are detected from behindthe east limb. In particular, the five circled three-spacecraft events (22 Septemberand 3 November 2011, 7 March and 23 July 2012, and 26 December 2013) hadintensities > 10−2 (MeV s cm2 sr)−1 when observed by a spacecraft for whichthe event was behind the east limb. These were among the most intense eventswhen observed by the best-connected spacecraft (also circled), suggesting thatthe intrinsic intensity of these events plays a role in their visibility at poorly-connected spacecraft. On the other hand, it is also evident that three-spacecraftevents have a wide range of intensities at well-connected longitudes and are notjust the most intense. In particular, there are events ∼ 3 orders of magnitudesmaller in intensity than the circled events that are also observed at all threespacecraft. An example is the 22 December 2009 event (Figure 13). The solarevent, associated with a C7.2 flare at W46o, was well-connected to the Earth,although interestingly, the proton event onset appears to have been more prompt



10-5

10-4

10-3

10-2STEREO A 14-24 MeV Protons 24-41 MeV Protons

10-5

10-4

10-3

Inte

nsity (

Me

V s

cm

**2

sr)

**-1

10-5

10-4

10-3

2009

355

21

356

22

357

23

358

24

359

25

Dec

DOY



1.0 AU

STEREO ASTEREO B

Earth/SOHO

Figure 13. Observations of a weak three-spacecraft event, on 22 December 2009. The verticalline gives the onset time of type III radio emissions (04:50 UT).

at STEREO B, for which the event was just behind the west limb, than at theEarth (both STEREO B and SOHO also detected electrons from this event).Type III radio emissions were observed at WIND and STEREOs A and B, andwere occulted at STEREO B as expected. No type II emissions were reported.LASCO observed a slow (318 km s−1), 47o width CME. Though the spacecraftseparation was clearly less than in other three-spacecraft events discussed above,and the event location was reasonably favorable for all three spacecraft, never-theless this event illustrates that even relatively weak events can be detectedover a wide region of the inner heliosphere.

Figure 14 summarizes several relationships between the ∼ 25 MeV protonpeak intensity and properties of the associated solar events for events in Decem-ber 2009–December 2012. Tables 1–7 indicate that every > 25 MeV proton event



0 1000 2000 3000

10-4

10-3

10-2

10-1

100

101

102

25 M

eV

Pro

ton I

nte

nsity

LASCO CME Speed (km/s)

cc=0.662

1 S/C

2 S/C

3 S/C

10-6 10-5 10-4 10-3 10-2

10-4

10-3

10-2

10-1

100

101

25 M

eV

Pro

ton Inte

nsity (

Eart

h)

GOES X-ray Intensity

cc = 0.711

B C M X

0 1000 2000 3000

10-4

10-3

10-2

10-1

100

101

102

25 M

eV

Pro

ton Inte

nsity


cc=0.712

IP Type II

Type II

No Type II

0 1 2 3 40

10

20

30

40

50

60

70

80

90

% o

f E

vents

No of S/C That Detect Event

No type II

Other type II

IP type II

Figure 14. The top-left panel shows the ∼ 25 MeV proton intensity versus the speed of theassociated LASCO CME, for cases in December 2009–December 2012 where the solar event ison the western front side hemisphere relative to the observing spacecraft. The symbol indicateswhether a particle event was observed at one, two or three spacecraft. The upper right panelshows the proton intensity at Earth plotted vs. the GOES soft X-ray flare intensity, for westernhemisphere events. The bottom left panel shows the proton intensity vs. CME speed (as in theupper left, but including a few events for which the number of spacecraft is uncertain). Greencrosses indicate that no WAVES type II radio emissions were reported, blue squares that “IP”type II emissions defined here as extending below 1 MHz were observed, and red circles thattype II emissions were detected but these did not extend below 1 MHz. The bottom right panelshows the percentage of events detected by one, two or three spacecraft that were accompaniedby no type II radio emissions, or by non-IP or IP type II emissions.

in this period had an associated CME, as also noted by Cane, Richardson andvon Rosenvinge (2010a) for similar events in Cycle 23. (A few late 2013 events donot have an assigned CME in the table, but only preliminary CME identificationsare available at the time of writing.) The top left panel of Figure 14 shows theproton intensity plotted against the speed of the LASCO CME for cases wherethe solar event is on the western front side hemisphere (i.e., W00–90o) relativeto the observing spacecraft (STEREO A, B, or SOHO). The well-establishedcorrelation between proton intensity and CME speed (e.g., Kahler, Hildner andVan Hollebeke, 1978; Kahler et al., 1984; Kahler, 2001; Cane, Richardson andvon Rosenvinge, 2010a, and references therein) is evident. The symbol type



indicates whether the proton event was detected at only one, only two or atall three spacecraft. Single-spacecraft events tend to be associated with slowerCMEs (below ∼ 1000 km s−1), while the fastest CMEs in this sample tend tobe associated with three-spacecraft events. Two-spacecraft events tend to haveintermediate speeds. However, there are overlaps between the distributions. Inparticular, the three-spacecraft events extend to slower CMEs, even below 1000km s−1, so a fast CME is evidently not a requisite for an SEP event that isextended in longitude.

CME width may also be correlated with proton intensity (e.g., Cane, Richard-son and von Rosenvinge, 2010a) but we have not examined this here for severalreasons. As is evident from Tables 1–7, around half (52%) of the events (toMay 2013) were associated with “full halo” CMEs (360o width) in the CDAWLASCO CME catalog, i.e., the ejection surrounds the coronagraph occultingdisc. However, as discussed in Cane, Richardson and von Rosenvinge (2010a), inmany cases, such halo CMEs are highly asymmetric. In that paper, halo CMEwidths were reassessed by the authors, with only symmetric halo CMEs retaininga width of 360o. While we could do this for the current set of events, the resultingCME widths would still include projection effects. These could be reduced byobtaining the CME width (and also speed) from the spacecraft (STEREO orSOHO) that is closest to quadrature relative to the solar event. However, thiswill be the focus of a future study.

The top right panel of Figure 14 shows the 25 MeV proton intensity plotted vs.the GOES soft X-ray flare peak intensity, for events on the western hemisphere asobserved from Earth. The proton and X-ray flare intensities are also correlated,with a correlation coefficient that is comparable to that for the proton intensitycorrelation with CME speed, as also found for the 25 MeV proton events in Cycle23 by Cane, Richardson and von Rosenvinge (2010a). On the other hand, in thediscussion above of the circumstances of the event in Figure 6, it was noted thatan earlier X1.9 flare was not accompanied by an SEP event. Including all 25 MeVproton events in this study for which an X-ray flare is listed in Tables 1–7, 9(7%) of these flares were B class, 47 (37%) were C class, 50 (40%) were M class,and 20 (16%) were X class. Thus, 77% of these proton events were associatedwith C or M class flares.

In Tables 1–7, 47% of the events (up to May 2013) do not have possible typeII emission observed by WIND/WAVES or SWAVES, 33% are accompanied byIP type II events, defined here as emission that extends below 1 MHz, while theremaining 20% have type II emissions that do not extend below this frequency.The bottom left panel of Figure 14 shows the proton intensity vs. CME speed,again for western hemisphere events at STEREO A, B, or SOHO, but in thiscase, the symbol indicates whether or not type II emission was reported, andwhether this could be classified as IP type II. The absence of type II emissionis evidently associated predominantly with low intensity proton events (below∼ 10−2 (MeV s cm2 sr)−1) with CME speeds below ∼ 1000 km s−1. Events withtype II emission are fairly well separated into more intense events with fasterCMEs that have IP type II emission and those with other type II emission thathave weak particle events and slower CMEs, with a distribution that significantlyoverlaps with the events without type II emission. Comparison with the top left



panel suggests that there might be some overlap between the events withouttype II emission and events detected at only one spacecraft, and similarly be-tween three-spacecraft events and those with IP type II emission. The bottomright panel in Figure 14 examines this further, showing the percentage of eventsdetected by one, two or three spacecraft that were or were not accompanied bytype II or IP type II emissions observed by WIND/WAVES or SWAVES. Theresults indicate that around three-quarters of single-spacecraft events were notaccompanied by type II emission, compared with around 50% of two-spacecraftevents, and around 20% of three-spacecraft events. Thus, the single-spacecraftevents do indeed tend to lack reported type II emission. On the other hand, thefraction with IP type II emission rises from around 10% for one-spacecraft eventsto around 30% for two-spacecraft events, and around 60% for three-spacecraftevents. The fraction with ‘other” type II emissions has little dependence on thenumber of detecting spacecraft and remains at ∼ 15–20% In the conventionalpicture that type II emissions are indicative of the formation of shocks, then apossible interpretation is that shocks are not present in the majority of single-spacecraft events, but are more likely to be present in multiple-spacecraft events.Alternatively, it is possible that such emissions may present in single-spacecraftevents but are too weak to detect.

Examining the presence of type III radio emissions, 92% of the events inTables 1–7 were accompanied by type III radio emissions, consistent with thehigh association between type III radio emission and SEP events noted byCane, Erickson and Prestage (2002) and Cane, Richardson and von Rosenvinge(2010a). A possible interpretation of events without type III emissions is thatthese do not involve any “flare-accelerated” particles, only particles acceleratedby the CME-driven shock, but a more detailed study of the individual events(such as 24 February 2012, which is associated with an erupting filament; seealso Cane, Kahler and Sheeley, 1986) is required to investigate this possibility.

3.3. Longitudinal dependence of three-spacecraft SEP events

3.3.1. Electron and proton delays to onset at 1 AU

We now combine observations of the three-spacecraft SEP events in December2009 to December 2012 to summarize the longitudinal dependence of several oftheir properties based on the ensemble of events. The top left panel of Figure 15shows the log of the ∼ 0.7–4 MeV electron onset delay (when this can be inferred)plotted vs. solar event longitude with respect to the observing spacecraft. Theelectron onset time is estimated from inspection of intensity–time profiles usingaveraging intervals of 1 minute to 1 hour (for weak events with extended delays).The errors indicate the estimated accuracy to which the time when the intensitysignificantly increases above fluctuations in the pre-event background can beinferred. The onset delays are measured relative to the start of the SWAVESor WIND/WAVES type III emissions as estimated from daily summary plots,since such emissions are observed for most events. Otherwise, the peak soft X-raytime is used. Symbols indicate observations of different events. Minimum delaysare associated with western hemisphere events that are expected to be better



-100 0 100

68

10

20

40

6080

100

200

400

600800

e-

De

lay(m

ins)

Longitude of Flare wrt S/C (deg.)

-100 0 100

68

10

20

40

6080

100

200

400

600800

e-

De

lay(m

ins)

Longitude of Field Footpoint wrt Flare (deg.)

0 100

68

10

20

40

6080

100

200

400

600800

e-

De

lay(m

ins)

|Longitude of Field Footpoint wrt Flare (deg.)|

cc = 0.684ln(y)=a*x+ba= 0.0153 +/- 0.0019b= 2.94 +/- 0.18

0 1000

100

200

300

400

500

600

700

800

900

Ele

ctr

on

So

urc

e S

pe

ed

(km

/s)@

1 R

s

|Longitude of Field Footpoint wrt. Flare (deg.)|

Figure 15. Summary of the ∼ 0.7–4 MeV electron onset delay versus solar event locationwith respect to the observing spacecraft (top left) and versus the angle between the footpointof the field line passing the spacecraft and the solar event (connection angle; top right), for thethree-spacecraft events in December 2009–December 2013. Symbols indicate different events.Delays are typically shortest for western hemisphere events when connection to the event isfavorable. The fits to events with positive or negative connection angles suggest that delaysare ∼symmetric to the east and west of the solar event. The bottom left panel shows the delayvs. the absolute connection angle. The delay at zero connection angle is ∼ 19 ± 3 minutes.The log-linear fit, if interpreted as increasing delays due to an electron “source” spreadingin longitude at the surface of the Sun suggests propagation speeds in longitude shown in thebottom right that fall with distance from the event, from ∼ 700 km s−1 near the event to∼ 130 km s−1 at 180o from the event.

connected by the IMF to the observing spacecraft. To examine the influence ofconnection, the top right panel shows the same observations plotted versus the“connection angle” (CA) between footpoint at the Sun of the nominal spiral fieldline passing the spacecraft at the time of the event, calculated using the observedsolar wind speed at the spacecraft (Nolte and Roeloff, 1973) and the longitudeof the solar event. A positive (negative) CA means that the footpoint lies west(east) of the solar event. (Note that this is the opposite convention to that usedby Lario et al., 2013 in their study of SEP longitudinal intensity dependence thatincludes some of the same events.) Electron onset delays tend to be shortest forwell-connected events. The fits to events with positive and negative connection



-100 0 10010

20

40

6080

100

200

400

600800

1000

P (

14-2

4 M

eV

) D

ela

y(m

ins)


-100 0 10010

20

40

6080

100

200

400

600800

1000

P (

14-2

4 M

eV

) D

ela

y(m

ins)


0 10010

20

40

6080

100

200

400

600800

1000

P (

14-2

4 M

eV

) D

ela

y(m

ins)


cc = 0.841ln(y)=a*x+b; Wtda= 0.0154 +/- 0.0011b= 3.98 +/- 0.095

0 1000

100

200

300

400

500

600

700

800

900

Pro

ton S

ourc

e S

peed (

km

/s)@

1 R

s

|Longitude of Field Footpoint wrt. Flare (deg.)|

Figure 16. Summary plots of 14–24 MeV proton onset delays, in a similar format to Figure 15.Onset delays are again typically shortest for western hemisphere events when connection tothe event is favorable. The fits to events with positive or negative connection angles (top left)suggest some east–west asymmetry in the proton onset delay. However, plotting the delays vs.| CA | (lower left) indicates a typical delay at zero connection angle of ∼ 54 minutes. The slopeof the log-linear fit is nearly identical to that for electrons. Thus, the ratio of the proton andelectron onset delays is ∼constant at all connection angles. The source expansion speeds areslower than for electrons, ranging from ∼ 240 km s−1 near the event to 45 km s−1 at ∼ 180o

from the event.

angles suggest that typical delays are approximately symmetric about the eventlocation. Thus, in the bottom left panel, the electron onset delays are plotted

versus | CA |. The fit indicates “typical” electron onset delays at 1 AU of ∼ 19minutes for CA = 0o increasing to ∼ 300 minutes (∼ 5 hours) for CA = 180o,though longer delays are found in some events.

A simple interpretation of the results is that the onset delay at 1 AU is givenby 4to = tI + t(φ), where tI includes the time for a particle to move from the

Sun to the spacecraft, assumed to be approximately constant for each event,plus any other “∼event-independent” delays relative to type III burst onset (for

example in the acceleration process). The typical delay at zero connection anglegives an estimate of the typical value of tI . The second term is an additionaldelay which increases with increasing connection angle φ at a typical rate that



is indicated by the fitted line in the lower left panel of Figure 15. This delaymight be ascribed to a “source” moving away from the solar event along thesurface of the Sun (or at some other height) which injects particles onto the fieldline foot points, after which they travel out to 1 AU. The slope of the fit in thelower left panel of Figure 15, which gives the rate of change in the delay as afunction of connection angle, can be expressed as the “propagation speed” of theelectron source shown in the bottom right panel of Figure 15 versus longitudefrom the event. The speed is higher closer to the event because the rate of changein the delay time with connection angle is smaller. It is also assumed that thesource is moving on the solar surface. The results suggest a deceleration from∼ 700 km s−1 near the event to ∼ 130 km s−1 at 180o from the event. For a sourceabove the solar surface, the speeds would be multiplied by the source distance(in Rs) from the center of the Sun. In practise, the source motion may be morecomplex (for example, it may lie on an expanding three-dimensional shock) butthis simple scenario at least gives an idea of the “propagation speeds” that mightbe implied by the observations with minimal interpretation. We also note thatother studies (e.g., Vainio et al., 2013) have used velocity dispersion analysiswith observations of particles over a range of energies/speeds to infer the solarrelease times of the SEPs. We have not made such an analysis in this paper sincewe are considering only two proton and one electron energy ranges, which areinsufficient for a reliable dispersion analysis, and are focusing on the arrival timeat 1 AU, which is most relevant for space weather applications, rather than onthe detailed timing of particle acceleration at the Sun.

Figure 16 shows similar results for the 14–24 MeV proton onset delay. Thetop-left panel shows that, as for electrons, minimum delays occur for well-connected events on the western hemisphere relative to the observing spacecraft.The weighted fits in the top-right panel suggest an east-west asymmetry in theproton delays, with shorter delays for connection to the east of the event forlarge connection angles. However, a non-weighted fit is more symmetric. Thebottom-left panel shows the correlation between the log of the proton onsetdelay and | CA |. The fit indicates that the typical delay at zero CA is 54 ± 5minutes, increasing to ∼ 860 ± 150 minutes at | CA |= 180o. Interestingly, theslopes of the fits in the bottom left panels of Figures 15 and 16 are essentiallyidentical (0.0153±0.0019 compared to 0.0154±0.0011) indicating that the ratioof the proton and electron onset delays derived from the fits is independent ofconnection angle, with a value of ∼ 2.8 that corresponds to the ratio of theproton and electron onset delays at CA = 0o. The ratio of electron and protononset delays will be discussed further below; See also the Helios spacecraft resultsof Kallenrode (1993) which also indicate longer onset delays for protons than forelectrons. The bottom-right panel of Figure 16 shows the proton “source propa-gation speed” inferred from the fit in the bottom-left panel. As for electrons, thesource speed decreases with increasing longitude from the event, but the speedsare lower, from ∼ 240 km s−1 close to the event, to ∼ 45 km s−1 at CA=180o,reflecting the longer proton onset delays.

Figure 17 examines whether there is any evidence of a correlation between the“source propagation speeds” at zero CA, inferred from fits to | CA | versus thelog of the electron or proton onset delay for individual events, and the speed of



1000 2000 3000

500

1000

1500

2000

Ele

ctr

on S

ourc

e S

peed, C

A=

0 (

km

/s)


cc = 0.016

1000 2000 3000

500

1000

1500

2000

Pro

ton S

ourc

e S

peed, C

A=

0 (

km

/s)


cc = -0.147

Figure 17. 0.7–4 MeV Electron (left) and 14–24 MeV proton “source propagation speeds” atzero CA inferred from | CA | vs. onset delays for individual events, plotted against the LASCOCME speed. The blue circle in the right-hand panel highlights the 3 November 2011 event.

6 810 20 40 60 100 200 400 800

20

40

60

80100

200

400

600

8001000

P (

14

-24

Me

V)

On

se

t D

ela

y(m

ins)

e- Onset Delay(mins)

cc = 0.934

ln(y)=a*ln(x)+b

a= 0.667 +/- 0.031

b= 2.3 +/- 0.13

6 810 20 40 60 100 200 400 600

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

P O

nse

t D

ela

y/e

- O

nse

t D

ela

y

e- Onset Delay(mins)

Figure 18. Left: Correlation of the electron and proton onset delays for individual obser-vations of the three-spacecraft events. Right: Ratio of proton to electron onset delays as afunction of electron onset delay.

the associated LASCO CME. The results indicate that there is little correlation,suggesting that the CME speed is not a good predictor of how rapidly particlesspread out in longitude from the solar event, at least in this simple scenario. Notethat the event with the outstanding high proton source speed (but not CMEspeed) circled in the right-hand panel is the 3 November 2011 event discussedabove.

The left-hand panel of Figure 18 shows the log of the 14–24 MeV proton onsetdelay plotted against the log of the 0.7–4 MeV electron onset delay for the sameevent/spacecraft. These are highly correlated (cc = 0.934). However, the bestfit indicates that 4toH ∼ 9.9 4 t0.67

oe , implying that the 4toH/ 4 toe ratio fallswith increasing onset delay, rather than the simple 4toH ∼ 2.84 toe suggestedby Figures 15 and 16. The right-hand panel of Figure 18 shows the ratio ofthe proton and electron onset delays for individual observations plotted againstelectron onset delay, illustrating the decline in the ratio with increasing delay,



-100 0 100

0.60.8

1

2

4

68

10

20

40

6080

e-

Pe

ak D

ela

y (

hrs

)


-100 0 100

0.60.8

1

2

4

68

10

20

40

6080

e-

Pe

ak D

ela

y (

hrs

)


0 100

0.60.8

1

2

4

68

10

20

40

6080

e-

Pe

ak D

ela

y (

hrs

)


cc = 0.543ln(y)=a*x+ba= 0.0118 +/- 0.0021b= 1.29 +/- 0.188

Figure 19. Summary of ∼ 0.7–4 MeV electron peak delays vs. flare longitude, connectionangle, and | CA |.

from ∼ 5 for minimal delays to ∼ 1–3 for long delays. Investigating the possibleorigin of this change in the ratio requires a more complete analysis of the onsettimes in individual events than is made here. (A possible scenario, for example,is that the initial proton acceleration/injection is delayed relative to electrons.)Nevertheless, the main conclusion to be drawn here is that overall, proton andelectron onset delays in three-S/C events, both as a function of CA and fromevent to event, are highly correlated. We note that Posner (2007) proposed usingobservations of near-relative electron onsets to give a ∼ 1 hour or more warningof the arrival of more hazardous protons and heavier ions with energies of 10s ofMeV. The results presented here provide further justification for this technique.

3.3.2. Delays to peak intensity

We now consider delays to peak particle intensity. Figure 19 summarizes thedelay to peak intensity versus event location with respect to the observing space-craft, CA, and | CA | for ∼ 0.7–4 MeV electrons. As for the onset delays, delaysto electron peak intensity are shortest for western hemisphere/well connectedevents. The fits in the top-right panel suggest a slight asymmetry, with shorter



-100 0 100

1

2

4

6

810

20

40

60

80

14

-24

Me

V H

Pe

ak D

ela

y (

hr)


-100 0 100

1

2

4

6

810

20

40

60

80

14

-24

Me

V H

Pe

ak D

ela

y (

hr)


0 100

1

2

4

6

810

20

40

60

80

14

-24

Me

V H

Pe

ak D

ela

y (

hr)


cc = 0.582ln(y)=a*x+ba= 0.00879 +/- 0.0014

b= 1.99 +/- 0.129

0 100

1

2

4

6

810

20

40

60

80

24

-41

Me

V H

Pe

ak D

ela

y (

hrs

)


cc = 0.613ln(y)=a*x+ba= 0.0109 +/- 0.0016

b= 1.59 +/- 0.145

Figure 20. Summary of 14–24 MeV proton peak delays vs. flare longitude, connection angle,and | CA |. The bottom right panel shows 24–41 MeV proton peak delays vs. | CA |.

delays for cases where the field line connects east of the event, though there isconsiderable scatter. The fit in the bottom panel indicates a typical delay toelectron peak of ∼ 3.6 hours at zero connection angle, increasing to ∼ 30 hoursat 180o connection angle.

Figure 20 illustrates similar results for 14–24 MeV protons, which again showthe shortest delays for western hemisphere/well-connected events, and approxi-mately symmetrical fits around zero connection angle. Note that, although thefocus of our analysis is on the early stages (∼ first day) of SEP events, thereare cases, typically for poorly-connected events, where particle intensities peakafter more extended delays. The fit in the bottom left panel indicates a typical14–24 MeV proton delay to peak of ∼ 7.3 hours at zero connection angle, and∼ 40 hours at 180o connection angle. The bottom right panel shows similarresults for 24–41 MeV protons; the fit indicates slightly shorter typical delaysof ∼ 5 hours at zero connection angle, and ∼ 30 hours at 180o. The variationof proton peak delay with event longitude is very consistent with that obtainedby Van Hollebeke, Ma Sung and McDonald (1975) for 125 SEP events observedat the Earth by IMPs 4 or 5, but by using the STEREO observations, it ispossible unambiguously to identify events with far-side sources, and observe



1 2 4 6 8 10 20 40 60 80

1

2

4

6

810

20

40

60

80

14

-24

Me

V H

Pe

ak D

ela

y (

hr)

e- Peak Delay (hrs)

cc = 0.861ln(y)=a*ln(x)+ba= 0.609 +/- 0.043

b= 1.31 +/- 0.104

1 2 4 6 8 10 20 40 60 800

1.0

2.0

3.0

4.0

5.0

6.0

7.0

H P

ea

k D

ela

y/e

- P

ea

k D

ela

y

e- Peak Delay (hrs)

Figure 21. Left: Correlation of the electron and proton delays to peak for individual ob-servations of the three-spacecraft events. Right: Ratio of proton to electron peak delays as afunction of electron onset delay.

events at multiple locations, enabling more events at large connection angles tobe included.

Figure 21 shows the correlation between electron and 14–24 MeV protondelays to peak intensity. The proton delay is related to the electron delay by4tpH = 3.74t0.61

pe . The right-hand panel shows the ratio of the proton andelectron delays as a function of electron delay. The ratio decreases from ∼ 4 forthe shortest delays, to ∼ 1, i.e., similar electron and proton delays to peak, forelectron delays above ∼ 10 hours.

We finally compare the electron or proton onset and peak delays. Since bothincrease with connection angle, we might expect some correlation as is shown tobe the case in Figure 22. Although there is an overall correlation, the electronresults suggest that below electron onset delays of ∼ 1 hour, the delays to peakare highly variable, ranging from ∼ 1 to 20 hours. The events to the left of themain distribution have an initial rapid onset followed by a slow rise to peakintensity. Investigation of reasons for these differences in the early developmentof the events, which may involve the evolution of the event at the Sun, inter-planetary propagation conditions, and the influence of local structures at theobserving spacecraft, is beyond the scope of this paper.

3.3.3. Longitudinal dependence of SEP intensity

To investigate the longitudinal dependence of SEP intensity, we again focus onthe three-spacecraft events in December 2009–December 2012. The top row ofFigure 23 shows Gaussian fits I(φ) = Ioe

−(φ−φo)2/2σ2

to background-corrected14–24 MeV proton intensities plotted versus connection angle (φ), for thoseevents where such a fit can be made. The left and right panels show events up toor after October 2011, respectively, to reduce the number of superposed graphs.The peak intensities (Io) of the Gaussian fits cover ∼ 5 orders of magnitude,again illustrating how the three-spacecraft events cover a wide range of intensi-



6 810 20 40 6080100 200 400 800

0.60.8

1

2

4

68

10

20

40

6080

e-

Pe

ak D

ela

y (

hrs

)

e- Onset Delay (mins)

cc = 0.755ln(y)=a*ln(x)+ba= 0.742 +/- 0.077b= -0.906 +/- 0.325

40 60 80100 200 400 600 1000

1

2

4

6

810

20

40

60

80

14

-24

Me

V H

Pe

ak D

ela

y (

hr)

14-24 MeV H Onset Delay (mins)

cc =0 .711ln(y)=a*ln(x)+ba = 0.629 +/- 0.0718

b = -0.571 +/- 0.371

Figure 22. Left: Electron peak delay plotted against electron onset delay. Right: 14–24 MeVproton peak delays plotted against onset delays.

-100 0 100

10-3

10-2

10-1

100

101

102

14

-24

Me

V P

roto

n I

nte

nsity

Longitude of Field Line Footpoint wrt. Flare (deg.)

-100 0 100

10-3

10-2

10-1

100

101

102


-100 0 100

10-3

10-2

10-1

100

101

102

-100 0 100

10-3

10-2

10-1

100

101

102

-100 0 100

100

101

102

103

104

0.7

-4 M

eV

e-

Inte

nsity


-100 0 100

100

101

102

103

104

0.7

-4 M

eV

e-

inte

nsity


-100 0 100

100

101

102

103

104

-100 0 100

100

101

102

103

104

Figure 23. Top row: Gaussian fits to the peak 14–24 MeV proton intensity((MeV s cm2 sr)−1) for three-spacecraft events in (left) December 2009–October 2011,and (right) November 2011–October 2012. In some cases, two estimates of the intensity atSOHO (from ERNE and EPHIN) are shown. A background of 10−4 (MeV s cm2 sr)−1

has been removed. Bottom row: Gaussian fits to the peak ∼ 0.3–4 MeV electron intensity((MeV s cm2 sr)−1) for three-spacecraft events in (left) December 2009–October 2011, and(right) November 2011–October 2012. STEREO HET intensities have been multiplied by thefactors discussed in Section 2.



ties. However, the weaker events, with Io < 10−1 (MeV s cm2 sr)−1 evidentlywere predominantly observed before October 2011.

The top left panel of Figure 24 shows how the connection angles of the peaksof the Gaussian fits (φo) tend to cluster around the longitude of the solar event.The average offset is 15.1o west, i.e., peak intensities, on average, are found onfield lines that connect ∼ 15o west of the solar event. However, the distributionis broad, with a standard deviation of 35.2o, so the results do not appear toindicate any consistent west/east bias. Uncertainty in the location of the solarevent, which may be extended and not accurately represented by the location ofthe related flare or other feature, and errors in inferring the spacecraft magneticconnection to the Sun, probably also contribute to the offset. The top right panelof Figure 24 shows the distribution of full widths at half maximum (FWHM,≈ 2.355σ) for 14–24 MeV protons. The average FWHM is 101.5o with a standarddeviation of 30.1o, corresponding to σ = 43 ± 13o. This is comparable to σ =45 ± 2o (σ = 43 ± 2o) found by Lario et al. (2013) for an independent (thoughoverlapping) sample of 25–53 MeV (15–40 MeV) proton events in 2009–2012,and the σ = 36 ± 2o obtained by Lario et al. (2006) for 27–37 MeV protonevents in Solar Cycle 21. Longitudinal SEP intensity dependences inferred frommulti-spacecraft observations were also reported by Kallenrode (1993) thoughGaussian fits were not made.

The bottom panels in Figures 23 and 24 show similar results for ∼ 0.7–4 MeV electrons for those events where such a fit can be made. The samesymbols are used for each event in the proton and electron plots. As for protons,the average footpoint location of the electron Gaussian maxima is displacedslightly to the west of the solar event by a similar angle (18.2o) though witha standard deviation of 28.9o. The electron FWHM is also similar to that forprotons (110 ± 32o), corresponding to σ = 47 ± 14o. This value is consistentwith the σ = 49 ± 2o (σ = 46 ± 2o) obtained by Lario et al. (2013) for 71–112 keV (0.7–3 MeV) electrons. The comparable σs for electrons and protonsimply that intensities decline with increasing longitude from the solar event atsimilar rates. We also note that Wiedenbeck et al. (2013) examined a 3He-richimpulsive event on 7 February 2010 (unrelated to the 25 MeV proton event onthis day in Table 1) that was observed at the two STEREO spacecraft and byACE when separated by 136o in longitude, and found that the ∼ 3 MeV/n3He fluence (event-integrated intensity) fell off with longitude with σ = 48o.These similar values of σ may suggest a common propagation scenario for 3Hein impulsive events and protons and electrons in the early stages of gradualSEP events. We also note that Kallenrode (1993) concluded that longitudinalintensity dependences are similar for “impulsive” events, defined in that studyby the X-ray flare duration (cf., Cane, McGuire and von Rosenvinge, 1986), andother events.

The left-hand panel of Figure 25 shows that proton and electron Gaussianpeaks in individual events tend to occur at similar connection angles – the dashedline indicates equality – while the best fit suggests a trend for the electron peakto be slightly closer to the event location. The proton and electron FWHMsare also correlated (right-hand panel of Figure 25), though the electron widthsgenerally slightly exceed the proton widths, typically by ∼ 10o.



-100 -50 0 50 1000

1

2

3

4

5

6

7

8

9

10

No. of E

vents

Connection Angle of Proton Peak (deg.)

Mean= 15.1 +/- 35.2

0 100 2000

1

2

3

4

5

6

7

8

9

10

No. of E

vents

14-24 MeV Proton FWHM (deg.)

Mean= 101.5 +/- 30.1

-100 -50 0 50 1000

1

2

3

4

5

6

7

8

No. of E

vents

Connection Angle of Electron Peak (deg.)

Mean= 18.2 +/- 28.9

0 100 2000

1

2

3

4

5

6

7

8

No. of E

vents

0.7-4 MeV Electron FWHM (deg.)

Mean= 110.4 +/- 31.7

Figure 24. Distributions of the connection angle of the maximum intensity, and full width halfmaximum, for proton (top row) and electron (bottom row) Gaussian fits to three-spacecraftevents in Figure 23.

-100 -50 0 50 100

-100

-80

-60

-40

-20

0

20

40

60

80

100

Connection A

ngle

of e-

Peak (

deg.)

Connection Angle of Proton Peak (deg.)

cc=0.932y=a*x+ba= 0.73 +/- 0.08b= 5.1 +/- 3.2

100 200

50

100

150

200

e-

FW

HM

(deg.)

Proton FWHM (deg.)

cc=0.870y=a*x+ba= 0.89 +/- 0.13b= 22.5 +/- 14.1

Figure 25. Correlation of proton and electron Gaussian peak connection angles (left) andFWHM (right) for three-spacecraft events. Symbols indicate events. The dashed lines indicateequal values for protons and electrons.



0 1000 2000 3000

10-3

10-2

10-1

100

101

102

14

-24

Me

V P

roto

n G

au

ssia

n H

eig

ht


cc = 0.771ln(y) = a*x+ba = 0.0036 +/- 0.0007b = -4.36 +/- 0.98

0 1000 2000 30000

50

100

150

200

14

-24

Me

V P

roto

n F

WH

M (

de

g.)


cc = 0.115y = a*x+ba = 0.0054 +/- 0.010b= 94.1 +/- 15.4

0 1000 2000 300010

0

101

102

103

104

0.7

-4 M

eV

Ele

ctr

on

Ga

ussia

n H

eig

ht


cc = 0.483ln(y)=a*x+ba = 0.00178 +/- 0.00086b = 1.88 +/- 1.4

0 1000 2000 30000

50

100

150

200

0.7

-4 M

eV

Ele

ctr

on

FW

HM

(d

eg

.)


cc = 0.470y = a*x+ba= 0.0277 +/- 0.0139b= 68.5 +/- 22.7

Figure 26. Correlation of LASCO CME speeds with proton (top row) and electron (bottomrow) Gaussian peak heights and FWHMs for three-spacecraft events. Symbols indicate events.

Figure 26 shows, for 14–24 MeV protons (top row) and 0.7–4 MeV protons(bottom row), the intensity of the peak of the Gaussian fit (left-hand panels)or FWHM (right-hand panels) plotted against the LASCO CME speed. As forthe observed proton intensities of western-hemisphere events in Figure 14, theproton Gaussian heights are correlated with the CME speed. The scatter in thepoints is reduced, as might be expected because variations due to connectionangle are not longer present. The electron peak intensity shows a weaker trendwith CME speed. The CME speed shows no (for protons) or little (for electrons)correlation with the FWHM. Thus, the observations do not appear to favor ascenario in which a faster CME results in a “wider” particle event that falls offless rapidly with longitude.

3.4. A “typical” event

The above observations suggest the following characteristics of a “typical” three-spacecraft SEP event. From the time of type III emissions onset, ∼ 0.7–4 MeVelectrons arrive at a spacecraft directly connected to the event typically in ∼ 19minutes, while 14–24 MeV protons do so after ∼ 54 minutes. The additional



10-3

10-2

10-1

100

101

102

10-3

10-2

10-1

100

101

102

Pre

dic

ted 1

4-2

4 M

eV

pro

ton Inte

nsity

Observed 25 MeV proton Intensity

cc=0.619ln(y)=a*ln(x)+ba = 0.453 +/- 0.025b = 0.454 +/- 0.147

Figure 27. Predicted 14–24 MeV proton intensity based on the relationshipbetween intensityand CME speed, and longitudinal dependence, for three-spacecraft events, plotted against theobserved intensity at 25 MeV for 532 SEP events observed at Earth and/or at STEREO A/Bin 1996–2013. The red line indicates equality.

proton delay may be largely accounted for by the difference in the particlespeeds along a nominal magnetic field line. For a spacecraft on a field line with aconnection angle φ(o) east or west of the solar event, particle arrival relative tothe type III burst is delayed by 4toe ≈ 19 exp(0.0153φ) minutes for 0.7–4 MeVelectrons (Figure 15) or 4toH ≈ 54 exp(0.0154φ) minutes for 14–24 MeV protons(Figure 16). The peak intensity is delayed by 4tpe ≈ 3.6 exp(0.0118φ) hoursfor 0.7–4 MeV electrons and 4tpH ≈ 7.3 exp(0.0088φ) hours for 14–24 MeVprotons. The peak intensity (not associated with passage of an interplanetaryshock) falls off with connection angle with FWHMs of ∼ 100o (σ ≈ 43o) forprotons and ∼ 110o (σ ≈ 47o) for electrons based on Gaussian fits. The peaksof the 14–24 MeV Gaussian fits are correlated with the CME speed V (km s−1)as Io ≈ 0.013 exp(0.0036V ) (MeV s cm2 sr)−1. (The observed particle intensityis also correlated with the soft X-ray flare intensity (Figure 14), but there areinsufficient events with Gaussian fits and associated X-ray observations to infer asimilar correlation.) Combining these results, the typical peak 14–24 MeV protonintensity (early in the event) at a spacecraft at 1 AU with a connection angleφ associated with a CME of speed V may be given by I(φ) (MeV s cm2 sr)−1

≈ 0.013 exp(0.0036V − φ2/2σ2)), σ = 43o. Figure 27 shows the predicted 14–24 MeV proton intensity based on this formula for 544 SEP events observed atthe Earth during 1996-2013 (including events from Cane, Richardson and vonRosenvinge, 2010a), or at STEREO A or B, plotted against the observed peak25 MeV proton intensity for these events. For simplicity, it is assumed for all



-100 0 100

10-3

10-2

10-1

100

101

102

103

104

(MeV

s c

m**

2 s

r)**

-1

Event Longitude(deg.)

Obs. ~25 MeV Proton Intensity

Pred. 14-24 MeV Proton Intensity

-100 0 100

10-3

10-2

10-1

100

101

102

103

104

Connection Angle(deg.)

Figure 28. Predicted 14–24 MeV proton intensity (red crosses) and observed ∼ 25 MeVproton intensity (blue circles) plotted against the solar event longitude relative to the observingspacecraft (left) and connection angle (right) for 544 SEP events observed at Earth and/or atSTEREO A/B in 1996–2013.

events that the solar wind speed is 450 km s−1 when calculating the connectionangle, and that peak SEP intensity at the Sun occurs at the longitude of thesolar event. The predicted and observed intensities do show a fair correlation(cc = 0.619), but there is considerable scatter, of around an order of magnitudeabove and below the fitted (black) line. Furthermore, the predicted intensitiesfor the weaker events tend to be overestimated - the red line indicates equality.This is also illustrated in Figure 28 which shows the predicted 14–24 MeV andobserved ∼ 25 MeV proton intensities for the 544 SEP events as a function ofevent longitude (left) and connection angle (right). The intensity variation withlongitude/connection angle for the largest events is reasonably well predicted, asmight be expected since the formula was developed using such events, suggestingthat it may have a potential for “forecasting” the intensities of the largest eventsthat are of particular interest for space weather. Nevertheless, it is clear that theformula, based only on CME speed and CA, overestimates the intensity fora large population of events. This is also evident from comparing the protonintensity–CME speed relationship for western hemisphere events in Figure 14and for the three-spacecraft event Gaussian peaks in Figure 26, noting the ∼ 2orders of magnitude difference in the intensities for events associated with slowerCMEs. Comparison of Figures 28 and 12 suggests that many of these weakerevents have characteristics similar to the “one” and “two” spacecraft eventsdiscussed above. However, many of the SEP events in Figure 28 are from thepre-STEREO era so the actual longitudinal extent of these events cannot bedetermined.

3.5. Implications for particle transport near the Sun and in the solar wind.

We now very briefly discuss the implications of the above observations for particletransport in longitude near the Sun and/or in the solar wind. There are a numberof processes that might be involved including:



• Multiple particle injections from “sympathetic” flares: For the three-spacecraftevents examined in this study, we find no clear case in which multiple solarevents in different active regions, closely-spaced in time, played a role inproducing what appears to be a single, longitudinally-extended particleevent.

• Cross-field transport/diffusion in the solar wind: Cross-field particle trans-port in the solar wind is generally expected to be a slow process comparedwith parallel transport, as indicated for example by clear influence of thespiral interplanetary magnetic field on particle transport evident in thisstudy, and by the particle intensity “drop outs” observed in some events byMazur et al. (2000). Giacalone and Kota (2012) have modeled the transportof particles in the solar wind from an impulsive event, motivated by theobservations of Wiedenbeck et al. (2013). In addition to cross-field diffusion,an important effect is corotation of the IMF which moves the particle pop-ulation towards the west, so that the particles can be detected, with somedelay, at spacecraft lying some distance west of the event. While this typeof process may contribute to the observation of particles from far-easternevents, cross-field transport in the solar wind is likely to be too slow toaccount for the “typical” onset delays at CA=180o of ∼ 5 hours for electronsand ∼ 14 hours for protons suggested by our analysis. The prompt onsetat widely-separated spacecraft on 3 November 2011 (Figure 7) providesa particular challenge for a scenario of longitudinal particle transport inthe solar wind. On the other hand, we note that Dresing et al. (2012)favor a role for strong perpendicular diffusion in the 17 January 2010 SEPevent based on the results of a 3D particle transport model. Such a processhas been also modeled recently by Marsh et al. (2013) and Lampa andKallenrode (2009), who conclude that it may contribute significantly toSEP longitudinal transport.

• Diffusion near/at the Sun/“coronal propagation”: This process has beenwidely studied in the past as a method of distributing solar particles inlongitude (e.g., Reinard and Wibberenz, 1974; Newkirk and Wentzel, 1978;Perez-Peraza, 1986), in particular before it was recognized that particleacceleration at extended CME-driven shocks could fulfill this role in largeSEP events (e.g., Cane, Reames and von Rosenvinge, 1988). It was alsosuggested by Wibberenz and Cane (2006) to be occurring during impulsiveelectron events. The tendency for east-west symmetry in the onset andpeak delays, and close correlation of the electron and proton delays mightbe suggestive of some type of diffusive process near to the Sun, at leastearly in gradual events.

• Acceleration by expansive interplanetary or coronal shocks: Cliver et al.(1995) noted that SEP events can turn on rapidly even at large connectionangles (for example, within two hours for ∼ 1 MeV electrons for connectionangles up to ∼ 150o; cf. the similar results in Figure 15). Considering suchevents in the context of shock acceleration, they concluded that accelerationat CME-driven shocks in the solar wind could not be responsible for theserapid distant onsets since such shocks appear to be limited to longitudinalextents of ∼ 180o. As an alternative, acceleration by essentially circum-solar



coronal shocks, with propagation speeds of ∼ 200− 500 km s−1 (based onelectron onset delays) was proposed. Cliver et al. (1995) did not examineproton onset delays in detail but suggested that differences in the protonand electron onset delays, as inferred in the present study, could pose aproblem for this scenario (as also noted by Kallenrode (1993) for coronalshocks in general), as could events with longer delays.

• Connection to an expanding three-dimensional CME shock: STEREO andSOHO observations now allow the three-dimensional structure of CMEsand their associated shocks to be inferred close to the Sun. Assuming thatSEPs are accelerated by CME-driven shocks, a reasonable scenario is thatSEPs are injected onto interplanetary field lines when the expanding shockintersects these field lines. For example, Rouillard et al. (2012) concludedthat the inter-spacecraft particle onset delays in the 21 March 2011 event(Figure 5) were consistent with connection to an expanding CME shockobserved by STEREO and SOHO. In this study we have identified furtherextended SEP events that may be used to establish whether this is a generalresult.

• Association with EUV waves: Since the discovery of EUV waves by the EITinstrument on SOHO, the possibility that particles are injected as such anexpanding EUV wave intercepts the foot points of magnetic field lines hasbeen considered (e.g., Bothmer et al., 1997; Krucker et al., 1999). Rouillardet al. (2012) noted that an EUV wave tracked the laterally-expanding flankof the 21 March 2011 CME, moving parallel to the surface of the Sunat ∼ 450 km s−1. AIA observations of EUV waves suggest initial speedsmay reach 600–2000 km s−1 but such waves decelerate rapidly, and typicalfinal speeds are 180–380 km s−1 (Patsourakos and Vourlidas, 2012, andreferences therein). Such speeds are reasonably comparable to the typical“source speeds” inferred in Figures 15, 16 and 17, in particular for protons,suggesting that EUV waves might be linked to the spread of SEPs in longi-tude. Nitta et al. (2013) have recently reported higher speeds (mean=644km s−1; median=607 km s−1) for large-scale coronal propagating frontsobserved by AIA, and noted that these speeds were poorly correlated withthe expansion speeds of the associated CMEs, which may be consistent withthe lack of correlation between source and CME speeds in Figure 17. Asfor the “coronal shock” scenario above, the different electron and protononset delays and inferred “source” speeds may pose a problem. In addition,an exceptionally fast and widespread EUV wave, extending to the footpoints of field lines connecting to STEREO B and Earth might have beenexpected to be associated with the 3 November 2011 event (Figure 6),but it appears that AIA did not observe such a wave (N. Nitta, privatecommunication, 2012; Nitta et al., 2013). Of the 171 fronts reported byNitta et al. (2013), 55 (32%) were associated with SEP events identifiedin this study, while of the 94 SEP events originating at E120o to W120o,the approximate range of the sources of the fronts reported by Nitta et al.(2013), 53 (56%) were associated with such a front. Of these 53 events,14 (26%) were single-spacecraft, 20 (38%) were two-spacecraft, and 12(23%) were three-spacecraft events (for the remaining events, the number



of spacecraft is uncertain), suggesting that observation of a propagatingfront is not strongly related to the observed extent of the SEP event.

• Particle transport to remote longitudes by large-scale magnetic loops: Re-constructions of coronal magnetic fields based on photospheric magneticfield observations (e.g., Schrijver et al., 2013) suggest the existence oflarge-scale magnetic loops that might guide SEPs to locations far from theoriginal solar event. Furthermore, modeling of the evolution of coronal mag-netic fields during a CME (Masson, Antiochos and DeVore, 2013) indicatesthat reconnection between closed and open field lines may lead to fieldconfigurations that facilitate particle transport to locations far removedfrom the original solar event. However, it is unclear whether the tendencyfor east-west symmetry in event delays and intensities and their orderingby connection angle, at least taken over the ensemble of events, would beconsistent with this scenario though it might play a role in particular events.

• Guidance by non-spiral interplanetary magnetic fields: Richardson, Caneand von Rosenvinge (1991) discussed cases where SEPs arrived promptlyfrom poorly-connected solar events by being guided within interplanetarycoronal mass ejections that happened to be passing the observing spacecraftat the times of the events. We have examined the solar wind present atthe three spacecraft at the times of the three-spacecraft SEP events, and ingeneral do not find evidence of such structures, though there are exceptionswhere local solar wind structure does appear to influence the developmentof SEP events at a particular spacecraft, leading to some of the outliersin the distributions discussed above. However discussion of such events isbeyond the scope of this paper.

4. Summary

We have used observations of protons and electrons from the HETs on theSTEREO A and B spacecraft, and from the ERNE and EPHIN instrumentson SOHO, to summarize the properties of solar particle events that included∼ 25 MeV protons since the beginning of the STEREO mission in October 2006until December 2013. The major conclusions are:

• Some 209 unique > 25 MeV proton events have been detected during thestudy period by the STEREO spacecraft and/or at Earth. All of the events,at least up to the end of the CDAW LASCO catalog at the time of writing(May 2013), were accompanied by CMEs.

• Considering events between December 2009 and December 2012, when thespacecraft were well separated, 36% were detected by only one spacecraft,34% by only two, and 17% by all three spacecraft, depending on factorssuch as whether particular spacecraft are or are not well connected to thesolar event by the nominal Parker spiral IMF, and the intensity of the event.Around 30% of the events were not detected at Earth.

• While the most intense events at well-connected spacecraft, and those as-sociated with the fastest CMEs, tend to be three-spacecraft events, the



intensity observed at the best-connected spacecraft, or the CME speed, forexample, are not reliable predictors of the visibility at the other spacecraft.

• 92% of the SEP events were accompanied by type III radio emissions ob-served by the WIND/WAVES or STEREO/SWAVES instruments, while53% had type II emissions, and 33% “IP” type II emissions extending below1 MHz.

• The percentage of events associated with type II emissions increases withthe number of spacecraft that detected the event, reaching 81% for three-spacecraft events, including ∼ 60% with IP type II emissions.

• Single-spacecraft events typically occur at well-connected longitudes, havepeak intensities at ∼ 25 MeV of < 10−2 (MeV s cm2 sr)−1, and are accom-panied by CMEs with speeds below 1000 km s−1. Type II emissions arereported in only 25% of cases.

• The availability of STEREO observations of the far side of the Sun confirmsthe results of previous studies using only observations at the Earth thataround a quarter of > 25 MeV proton events originate behind the westlimb with respect to the observer, and occasionally (∼ 4–8% of events)behind the east limb. Energetic particles are unambiguously observed fromsolar activity ∼ 180o in longitude from the observation point, and on fieldlines that have foot points at the Sun ∼ 180o from the solar event.

• SEP occurrence rates at Earth were reasonably comparable during therising phases of Cycle 23 and 24. However, the first event was detectedat Earth one year after smoothed sunspot minimum in Cycle 24 comparedwith a delay of only two months in Cycle 23. The first two years of Cycle24 were characterized by brief intervals with SEP activity followed by ∼sixmonth intervals of low activity that may be evidence of a “Rieger et al.(1984)-like” quasi-periodicity.

• The SEP rate during 2013 declined from that in 2011 and 2012, but showsevidence of an increase in the final months of 2013 that appears to beassociated with an increase in the number of southern hemisphere sunspots.The lower SEP rate in 2012–2013 relative to 2011 may be a manifestationof the “Gnevyshev gap” in energetic solar activity often observed duringsolar cycle maxima.

• Longitudinal-dependencies of the electron and proton peak intensities, anddelays to onset and peak have been examined for three-spacecraft events.Both delays to onset and delays to peak increase with increasing connectionangle between the solar event and the footpoint of magnetic field linespassing the observing spacecraft. The results suggest a close “coupling”between energetic protons and electrons such as correlated onset and peakdelays and similar intensity dependence with longitude.

• The 3 November 2011 event had an exceptionally rapid turn-on at all threespacecraft, with a difference in the ∼ 25 MeV proton onset times of only∼ 25 minutes. The CME speed however was un-exceptional (991 km s−1).

• A formula predicting the proton intensity at 14–24 MeV based on the CMEspeed and solar event location has been developed using three-spacecraftevent observations. While it appears to predict the intensity of the largest,most extended events reasonably well, it fails for a large population ofweaker than expected events.



• The tendency for east-west symmetry in the delays to onset, and peak,and the Gaussian fits to peak around zero connection longitude suggestthat electron and proton propagation in longitude away from the solarevent occurs somewhat symmetrically. This may be most consistent witha scenario involving a propagating shock or wave moving out from theevent, though different source propagation speeds inferred for electrons andprotons may be a complication, or some sort of diffusive process.

Acknowledgements We thank Janet Luhmann of the University of California, Berkeley,

for her steadfast support as the Principal Investigator of the STEREO/IMPACT investigation.

The work at GSFC, Caltech and JPL was supported by NASA (Caltech and JPL were funded

under subcontract SA2715-26309 from the University of California, Berkeley, under NASA

Contract NAS5-03131). The LASCO CME catalog at http://cdaw.gsfc.nasa.gov/CME list/is

generated and maintained at the CDAW Data Center by NASA and The Catholic University

of America in cooperation with the Naval Research Laboratory. We particularly appreci-

ate the efforts by S. Yashiro and colleagues to generate daily summary movies of SOHO

and STEREO observations (http://cdaw.gsfc.nasa.gov/stereo/daily movies/) which were very

helpful in verifying the solar sources of the SEP events. The CACTUS CME catalog is main-

tained by the Solar Influences Data Analysis Center at the Royal Observatory of Belgium.

The SOHO/ERNE data were obtained from the Space Research Laboratory, University of

Turku (http://www.srl.utu.fi/erne data/). The SOHO/EPHIN project is supported under grant

No. 50 OC 1302 by the German Bundesminister fur Wirtschaft through the Deutsches Zentrum

fur Luft und Raumfahrt (DLR). SOHO is a project of international cooperation between ESA

and NASA.

References

Bazilevskaya, G. A., Makhmutova, V S., Sladkova, A. I.: 2006, Adv. Space Res. 38(3), 484.Bothmer, V., Posner, A., Kunow, H., Muller-Mellin, R., Heber, B., Pick, M., et al.: 1997,

Proc. 31st ESLAB Symp. “Correlated Phenomena at the Sun, in the Heliosphere, and inGeospace”, ESA SP-415, 207.

Brueckner, G. E., Howard, R. A., Koomen, M. J., Korendyke, C. M., Michels, D. J., Moses,J. D., et al.: 1995, Solar Phys. 162, 357.

Cane, H. V.: 1985, J. Geophys. Res. 90, 191.Cane, H. V., Erickson, W. C.: 2003, J. Geophys. Res. 108(A5), 1203,

doi:10.1029/2002JA009488.Cane, H. V., Erickson, W. C.: 2005, Astrophys. J. 623, 1180.Cane, H. V., Stone, R. G.: 1984, Astrophys. J. 282, 339.Cane, H. V., Kahler, S. W., Sheeley, N. R., Jr: 1986, J. Geophys. Res. 91, 13,321.Cane, H. V., McGuire, R. E., von Rosenvinge, T. T.: 1986, Astrophys. J. 301, 448.Cane, H. V., Reames, D. V., von Rosenvinge, T. T.: 1988, J. Geophys. Res. 93, 9555.Cane, H. V., Erickson, W. C., Prestage, N. P.: 2002, J. Geophys. Res. 107(A10), 1315,

doi:10.1029/2001JA000320.Cane, H. V., Richardson, I. G., von Rosenvinge, T. T.: 1998, Geophys. Res. Lett. 25, 4437.Cane, H. V., Richardson, I. G., von Rosenvinge, T. T.: 2010a, J. Geophys. Res. 115, A08101,

doi:10.1029/2009JA014848.Cane, H. V., Richardson, I. G., von Rosenvinge, T. T.: 2010b, Twelfth International Solar

Wind Conference, Maksimovic, M., et al. (eds), AIP Conf. Proc. 1216, 687.Cane, H. V., Stone, R. G., Fainberg, J., Steinberg, J. L., Hoang, S.: 1982, Solar Phys. 78, 187.Chowdhury, P., Choudhary, D. P., Gosain, S.: 2013, Astrophys. J. 768, 188, doi:10.1088/0004-

637X/768/2/188.Cliver, E. W., Kallenrode, M.-B.: 2001, Proc. 27th Int. Cosmic Ray Conf. 8, 3318.



Cliver, E. W., Kahler, S. W., Neidig, D. F., Cane, H. V., Richardson, I. G., Kallenrode, M.-B.,Wibberenz, G.: 1995, Proc. 24th Int. Cosmic Ray Conf. 4, 257.

Dalla, S., Balogh, A., Heber, B., Lopate, C.: 2001, J. Geophys. Res. 106, 5721.Delaboudiniere, J.-P., Artzner, G. E., Brunaud, J., Gabriel, A. H., Hochedez, J. F., Millier,

F.: 1995, Solar Phys. 162, 291.Dresing, N., Gomez-Herrero, R., Klassen, A., Heber, B., Kartavykh, Y., Droge, W.: 2012, Solar

Phys. 281, 281, DOI 10.1007/s11207-012-0049-y.Feminella, F., Storini, M.: 1997, Astron. Astrophys. 322, 311.Giacalone, J., Kota, J.: 2012, Astrophys. J. 751, L33.Gnevyshev, M. N.: 1967, Sol. Phys. 1 107.Gnevyshev, M. N.: 1977, Sol. Phys. 51, 175.Gopalswamy, N., Yashiro, S., Michalek, G., Kaiser, M. L., Howard, R. A., Reames, D. V., et

al.: 2002, Astrophys. J. (Lett.) 572, L103.Gopalswamy, N., Yashiro, S., Krucker, S., Stenborg, G., Howard, R. A.: 2004, J. Geophys. Res.

109(A12), doi:10.1029/2004JA010602.Howard, R., Moses, J. D., Vourlidas, A., Newmark, J. S., Socker, D. G., Plunkett, S. P., et al.:

2008, Space Sci. Rev. 136, 67.Kahler, S. W.: 2001, J. Geophys. Res. 110 A12S01.Kahler, S. W., Hildner, E., Van Hollebeke, M. A. I.: 1978, Solar Phys. 57, 429.Kahler, S. W., Sheeley, N. R., Jr., Howard, R. A., Michels, D. J., Koomen, M. J., McGuire,

R. E., et al.: 1984, J. Geophys. Res. 89, 9683, doi:10.1029/JA089iA11p09683.Kallenrode, M.-B.: 1993, J. Geophys. Res. 98, 5573.Krucker, S., Larson, D. E., Lin, R. P., Thompson, B. J.: 1999, Astrophys. J. 519, 864.Lampa, F., Kallenrode, M.-B.: 2009, Solar Phys. 260, 423, DOI 10.1007/s11207-009-9465-z.Lario, D., Kallenrode, M.-B., Decker, R. B., Roelof, E. C., Krimigis, S. M., Aran, A., Sanahuja,

B.: 2006, Astrophys. J. 653, 1531.Lario, D., Aran, A.,Gomez-Herrero, R., Dresing, N., Heber, B., Ho, G. C., Decker, R. B.,

Roelof, E. C.: 2013, Astrophys. J. 767, 41.Lean, J. L.: 1990, Astrophys. J. 363, 718.Lemen, J. R., Title, A. M., Akin, D. J., Boerner, P. F., Chou, C., Drake, J. F., et al.: 2012,

Solar Phys. 275, 17, doi: 10.1007/s11207-011-9776-8.Leske, R. A., Cohen, C. M. S., Dotson, B., Mewaldt, R. A., Cummings, A. C., Labrador, A. W.,

et al.: 2013, Thirteenth International Solar Wind Conference, AIP Conf. Proc. 1539, 227,doi: 10.1063/1.4811029.

Levy, E. H., Duggal, S. P., Pomerantz, M. A.: 1976, J. Geophys. Res. 81, 51.Li, G., Moore, R., Mewaldt, R. A., Zhao, L., Labrador, A. W.: 2012, Space Sci. Rev. 171, 141,

DOI 10.1007/s11214-011-9823-7.Marsh, M. S., Dalla, S., Kelly, J., Laitinen, T.: 2013, Astrophys. J 774, 4, doi:10.1088/0004-

637X/774/1/4.Masson, S., Antiochos, S. K., DeVore, C. R.: 2013, Astrophys. J 771, 82.Mazur, J. E., Mason, G. M., Dwyer, J. R., Giacalone, J., Jokipii, J. R., Stone, E. C.: 2000,

Astrophys. J. 532, L79.Moon, Y.-J., Choe, G. S., Park, Y. D., Wang, H., Gallagher, P. T., Chae, J., Yun, H. S., Goode,

P. R.: 2002, Astrophysical J. 574, 434.Muller-Mellin, R., Kunow, H., Fleissner, V., Pehlke, E., Rode, E.. Roschmann, N., et al.: 1995,

Solar Phys. 162, 483.Newkirk, G., Jr., Wentzel, D. G.: 1978, J. Geophys. Res. 83, 2009.Nitta, N. V., Schrijver, C. J., Title, A. M., Liu, W.: 2013, Astrophys. J. 776, 58,

doi:10.1088/0004-637X/776/1/58.Nolte, J. T., Roeloff, E. C.: 1973, Solar Phys. 33, 241.Norton, A. A., Gallagher, J. G.: 2010, Solar Phys. 261, 193.Patsourakos, S., Vourlidas, A.: 2012, Solar Phys. 281, 187.Perez-Peraza, J.: 1986, Space Sci. Rev. 44, 91.Pomerantz, M. A., Duggal, S. P.: 1974, J. Geophys. Res. 79, 913.Posner, A.: 2007, Space Weather 5, S05001, doi:10.1029/2006SW000268.Reinard, R., Wibberenz, G.: 1974, Solar Phys. 36, 473.Richardson, R. S.: 1936, Ann. Report Dir. Mt. Wilson Obs., 35, 871.Richardson, I. G., Cane, H. V.: 2005, Geophys. Res. Lett. 32, L02104.Richardson, I. G., Cane, H. V.: 2010, Solar Phys. 264, 189.Richardson, I. G., Cane, H. V.: 2012, J. Space Weather Space Climate 2, A02.Richardson, I. G., Cane, H. V., von Rosenvinge, T. T.: 1991, J. Geophys. Res. 96, 7853.



Rieger E., Share, G. H., Forrest, D. J., Kanbach, G., Reppin, C., Chupp, E. L.: 1984, Nature312, 623.

Rouillard, A. P., Sheeley, N. R., Jr., Tylka, A., Vourlidas, A., Ng, C. K., Rakowski, C., et al.:2012, Astrophys. J. 752, 44.

Schrijver, C. J., Title, A. M.: 2011, J. Geophys. Res. 116, A04108, doi:10.1029/2010JA016224.Schrijver, C. J., Title, A. M., Yeates, A. R., DeRosa, M. L.: 2013, Astrophys. J., 773, 93.Storini, M., Bazilevskaya, G. A., Fluckiger, E. O., Krainev, M. B., Makhmutov, V. S., Sladkova,

A. I.: 2003, Adv. Space Res. 31(4), 895.Torsti, J., Valtonen, E., Lumme, M., Peltonen, P., Eronen, T., Louhola, M., et al.: 1995, Solar

Phys. 162, 505.Vainio, R., Valtonen, E., Heber, B., Malandraki, O. E., Papaioannou, A., Klein, K.-L., et al.:

2013, J. Space Climate Space Weath. 3, A12.Van Hollebeke, M. A. I., Ma Sung, L. S., McDonald, F. B.: 1975, Solar Phys. 41, 189.von Rosenvinge, T. T., Reames, D. V., Baker, R., Hawk, J., Nolan, J. T. Ryan,L., et al.: 2008,

Space Sci. Rev. 136, 391, doi: 10.1007/s11214-007-9300-5.von Rosenvinge, T. T., Richardson, I. G., Reames, D. V., Cohen, C. M. S., Cummings, A. C.,

Leske, R. A., et al.: 2009, Solar Phys. 256, 443.Wibberenz, G., Cane, H. V.: 2006, Astrophys. J. 650, 1199.Wiedenbeck,M. E., Mason, G. M., Cohen, C. M. S., Nitta, N. V., Gomez-Herrero,R., Haggerty,

D. K.: 2013, Astrophys. J. 762, 54.Wulser, J.-P., Lemen, J. R.., Tarbell, T. D., Wolfson, C. J., Cannon, J. C.; Carpenter, B. A.,

Duncan, D. W., et al.: 2004, Proc. SPIE 5171, Telescopes and Instrumentation for SolarAstrophysics, 111, doi:10.1117/12.506877.

Yashiro, S., Gopalswamy, N., Michalek, G., St. Cyr, O. C., Plunkett, S. P., Rich, N. B., Howard,R. A.: 2004, J. Geophys. Res. 109, A07105, doi:10.1029/2003JA010282.

Zurbuchen, T. H., Richardson, I. G.: 2006, Space Science Rev., 123, 31, doi:10.1007/s11214-006-9010-4.


25 mev proton events observed by the high energy ... code 290-17, ... 101 102 stereo a 14-24 mev...

Documents